Compare commits

..

12 Commits

Author SHA1 Message Date
Andrew Kane
e4ac05f044 Limit max probes [skip ci] 2024-09-22 11:20:28 -07:00
Andrew Kane
25f97fd91d Updated cost estimation [skip ci] 2024-09-22 11:17:29 -07:00
Andrew Kane
cd3f9a38ae Added max_probes option [skip ci] 2024-09-21 20:19:26 -07:00
Andrew Kane
ff6267917e Improved approach [skip ci] 2024-09-21 19:55:55 -07:00
Andrew Kane
c950c5ffaa Merge branch 'master' into ivfflat-streaming 2024-09-21 19:26:12 -07:00
Andrew Kane
88f56dc234 Merge branch 'master' into ivfflat-streaming 2024-09-21 18:37:31 -07:00
Andrew Kane
e2ba6cf38f Updated comment [skip ci] 2024-09-20 21:33:46 -07:00
Andrew Kane
689f9c4659 Added cost estimation [skip ci] 2024-09-20 21:30:52 -07:00
Andrew Kane
88889f5a4c Fixed CI [skip ci] 2024-09-20 12:52:08 -07:00
Andrew Kane
79851729f1 Improved streaming test 2024-09-20 11:13:03 -07:00
Andrew Kane
3fd6a29c49 Added test for streaming recall [skip ci] 2024-09-20 10:50:53 -07:00
Andrew Kane
2c58804756 Added streaming option for IVFFlat [skip ci] 2024-09-20 10:20:40 -07:00
53 changed files with 630 additions and 3687 deletions

View File

@@ -49,7 +49,7 @@ jobs:
- postgres: 16
os: macos-14
- postgres: 14
os: macos-13
os: macos-12
steps:
- uses: actions/checkout@v4
- uses: ankane/setup-postgres@v1

View File

@@ -1,10 +1,6 @@
## 0.8.0 (unreleased)
- Added support for iterative index scans
- Added `intvec` type
- Added casts for arrays to `sparsevec`
- Improved cost estimation
- Improved performance of HNSW inserts and on-disk index builds
- Reduced memory usage for HNSW index scans
- Dropped support for Postgres 12

View File

@@ -1,4 +1,4 @@
ARG PG_MAJOR=17
ARG PG_MAJOR=16
FROM postgres:$PG_MAJOR
ARG PG_MAJOR

View File

@@ -4,8 +4,8 @@ EXTVERSION = 0.7.4
MODULE_big = vector
DATA = $(wildcard sql/*--*--*.sql)
DATA_built = sql/$(EXTENSION)--$(EXTVERSION).sql
OBJS = src/bitutils.o src/bitvec.o src/halfutils.o src/halfvec.o src/hnsw.o src/hnswbuild.o src/hnswinsert.o src/hnswscan.o src/hnswutils.o src/hnswvacuum.o src/intvec.o src/ivfbuild.o src/ivfflat.o src/ivfinsert.o src/ivfkmeans.o src/ivfscan.o src/ivfutils.o src/ivfvacuum.o src/sparsevec.o src/vector.o
HEADERS = src/halfvec.h src/intvec.h src/sparsevec.h src/vector.h
OBJS = src/bitutils.o src/bitvec.o src/halfutils.o src/halfvec.o src/hnsw.o src/hnswbuild.o src/hnswinsert.o src/hnswscan.o src/hnswutils.o src/hnswvacuum.o src/ivfbuild.o src/ivfflat.o src/ivfinsert.o src/ivfkmeans.o src/ivfscan.o src/ivfutils.o src/ivfvacuum.o src/sparsevec.o src/vector.o
HEADERS = src/halfvec.h src/sparsevec.h src/vector.h
TESTS = $(wildcard test/sql/*.sql)
REGRESS = $(patsubst test/sql/%.sql,%,$(TESTS))
@@ -66,7 +66,7 @@ dist:
git archive --format zip --prefix=$(EXTENSION)-$(EXTVERSION)/ --output dist/$(EXTENSION)-$(EXTVERSION).zip master
# for Docker
PG_MAJOR ?= 17
PG_MAJOR ?= 16
.PHONY: docker

View File

@@ -2,8 +2,8 @@ EXTENSION = vector
EXTVERSION = 0.7.4
DATA_built = sql\$(EXTENSION)--$(EXTVERSION).sql
OBJS = src\bitutils.obj src\bitvec.obj src\halfutils.obj src\halfvec.obj src\hnsw.obj src\hnswbuild.obj src\hnswinsert.obj src\hnswscan.obj src\hnswutils.obj src\hnswvacuum.obj src\intvec.obj src\ivfbuild.obj src\ivfflat.obj src\ivfinsert.obj src\ivfkmeans.obj src\ivfscan.obj src\ivfutils.obj src\ivfvacuum.obj src\sparsevec.obj src\vector.obj
HEADERS = src\halfvec.h src\intvec.h src\sparsevec.h src\vector.h
OBJS = src\bitutils.obj src\bitvec.obj src\halfutils.obj src\halfvec.obj src\hnsw.obj src\hnswbuild.obj src\hnswinsert.obj src\hnswscan.obj src\hnswutils.obj src\hnswvacuum.obj src\ivfbuild.obj src\ivfflat.obj src\ivfinsert.obj src\ivfkmeans.obj src\ivfscan.obj src\ivfutils.obj src\ivfvacuum.obj src\sparsevec.obj src\vector.obj
HEADERS = src\halfvec.h src\sparsevec.h src\vector.h
REGRESS = bit btree cast copy halfvec hnsw_bit hnsw_halfvec hnsw_sparsevec hnsw_vector ivfflat_bit ivfflat_halfvec ivfflat_vector sparsevec vector_type
REGRESS_OPTS = --inputdir=test --load-extension=$(EXTENSION)

View File

@@ -52,8 +52,6 @@ nmake /F Makefile.win
nmake /F Makefile.win install
```
Note: Postgres 17 is not supported yet due to an upstream issue
See the [installation notes](#installation-notes---windows) if you run into issues
You can also install it with [Docker](#docker) or [conda-forge](#conda-forge).
@@ -102,8 +100,6 @@ Or add a vector column to an existing table
ALTER TABLE items ADD COLUMN embedding vector(3);
```
Also supports [half-precision](#half-precision-vectors), [binary](#binary-vectors), and [sparse](#sparse-vectors) vectors
Insert vectors
```sql
@@ -149,8 +145,6 @@ Supported distance functions are:
- `<#>` - (negative) inner product
- `<=>` - cosine distance
- `<+>` - L1 distance (added in 0.7.0)
- `<~>` - Hamming distance (binary vectors, added in 0.7.0)
- `<%>` - Jaccard distance (binary vectors, added in 0.7.0)
Get the nearest neighbors to a row
@@ -223,7 +217,7 @@ L2 distance
CREATE INDEX ON items USING hnsw (embedding vector_l2_ops);
```
Note: Use `halfvec_l2_ops` for `halfvec`, `intvec_l2_ops` for `intvec`, and `sparsevec_l2_ops` for `sparsevec` (and similar with the other distance functions)
Note: Use `halfvec_l2_ops` for `halfvec` and `sparsevec_l2_ops` for `sparsevec` (and similar with the other distance functions)
Inner product
@@ -259,7 +253,6 @@ Supported types are:
- `vector` - up to 2,000 dimensions
- `halfvec` - up to 4,000 dimensions (added in 0.7.0)
- `intvec` - up to 8,000 dimensions (added in 0.8.0)
- `bit` - up to 64,000 dimensions (added in 0.7.0)
- `sparsevec` - up to 1,000 non-zero elements (added in 0.7.0)
@@ -478,16 +471,6 @@ Get the nearest neighbors
SELECT * FROM items ORDER BY embedding::halfvec(3) <-> '[1,2,3]' LIMIT 5;
```
## Integer Vectors
*Added in 0.8.0*
Use the `intvec` type to store 8-bit integer vectors
```sql
CREATE TABLE items (id bigserial PRIMARY KEY, embedding intvec(3));
```
## Binary Vectors
Use the `bit` type to store binary vectors ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/imagehash/example.py))
@@ -951,30 +934,6 @@ Function | Description | Added
avg(halfvec) → halfvec | average | 0.7.0
sum(halfvec) → halfvec | sum | 0.7.0
### Intvec Type
Each int vector takes `dimensions + 8` bytes of storage. Each element is a single byte signed integer. Int vectors can have up to 16,000 dimensions.
### Intvec Operators
Operator | Description | Added
--- | --- | ---
<-> | Euclidean distance | 0.8.0
<#> | negative inner product | 0.8.0
<=> | cosine distance | 0.8.0
<+> | taxicab distance | 0.8.0
### Intvec Functions
Function | Description | Added
--- | --- | ---
cosine_distance(intvec, intvec) → double precision | cosine distance | 0.8.0
inner_product(intvec, intvec) → double precision | inner product | 0.8.0
l1_distance(intvec, intvec) → double precision | taxicab distance | 0.8.0
l2_distance(intvec, intvec) → double precision | Euclidean distance | 0.8.0
l2_norm(intvec) → double precision | Euclidean norm | 0.8.0
vector_dims(intvec) → integer | number of dimensions | 0.8.0
### Bit Type
Each bit vector takes `dimensions / 8 + 8` bytes of storage. See the [Postgres docs](https://www.postgresql.org/docs/current/datatype-bit.html) for more info.
@@ -1024,7 +983,7 @@ l2_normalize(sparsevec) → sparsevec | Normalize with Euclidean norm | 0.7.0
If your machine has multiple Postgres installations, specify the path to [pg_config](https://www.postgresql.org/docs/current/app-pgconfig.html) with:
```sh
export PG_CONFIG=/Library/PostgreSQL/17/bin/pg_config
export PG_CONFIG=/Library/PostgreSQL/16/bin/pg_config
```
Then re-run the installation instructions (run `make clean` before `make` if needed). If `sudo` is needed for `make install`, use:
@@ -1035,11 +994,11 @@ sudo --preserve-env=PG_CONFIG make install
A few common paths on Mac are:
- EDB installer - `/Library/PostgreSQL/17/bin/pg_config`
- Homebrew (arm64) - `/opt/homebrew/opt/postgresql@17/bin/pg_config`
- Homebrew (x86-64) - `/usr/local/opt/postgresql@17/bin/pg_config`
- EDB installer - `/Library/PostgreSQL/16/bin/pg_config`
- Homebrew (arm64) - `/opt/homebrew/opt/postgresql@16/bin/pg_config`
- Homebrew (x86-64) - `/usr/local/opt/postgresql@16/bin/pg_config`
Note: Replace `17` with your Postgres server version
Note: Replace `16` with your Postgres server version
### Missing Header
@@ -1048,10 +1007,10 @@ If compilation fails with `fatal error: postgres.h: No such file or directory`,
For Ubuntu and Debian, use:
```sh
sudo apt install postgresql-server-dev-17
sudo apt install postgresql-server-dev-16
```
Note: Replace `17` with your Postgres server version
Note: Replace `16` with your Postgres server version
### Missing SDK
@@ -1084,17 +1043,17 @@ If installation fails with `Access is denied`, re-run the installation instructi
Get the [Docker image](https://hub.docker.com/r/pgvector/pgvector) with:
```sh
docker pull pgvector/pgvector:pg17
docker pull pgvector/pgvector:pg16
```
This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (replace `17` with your Postgres server version, and run it the same way).
This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (replace `16` with your Postgres server version, and run it the same way).
You can also build the image manually:
```sh
git clone --branch v0.7.4 https://github.com/pgvector/pgvector.git
cd pgvector
docker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector .
docker build --pull --build-arg PG_MAJOR=16 -t myuser/pgvector .
```
### Homebrew
@@ -1105,7 +1064,7 @@ With Homebrew Postgres, you can use:
brew install pgvector
```
Note: This only adds it to the `postgresql@17` and `postgresql@14` formulas
Note: This only adds it to the `postgresql@14` formula
### PGXN
@@ -1120,22 +1079,22 @@ pgxn install vector
Debian and Ubuntu packages are available from the [PostgreSQL APT Repository](https://wiki.postgresql.org/wiki/Apt). Follow the [setup instructions](https://wiki.postgresql.org/wiki/Apt#Quickstart) and run:
```sh
sudo apt install postgresql-17-pgvector
sudo apt install postgresql-16-pgvector
```
Note: Replace `17` with your Postgres server version
Note: Replace `16` with your Postgres server version
### Yum
RPM packages are available from the [PostgreSQL Yum Repository](https://yum.postgresql.org/). Follow the [setup instructions](https://www.postgresql.org/download/linux/redhat/) for your distribution and run:
```sh
sudo yum install pgvector_17
sudo yum install pgvector_16
# or
sudo dnf install pgvector_17
sudo dnf install pgvector_16
```
Note: Replace `17` with your Postgres server version
Note: Replace `16` with your Postgres server version
### pkg

View File

@@ -1,188 +1,6 @@
-- complain if script is sourced in psql, rather than via CREATE EXTENSION
\echo Use "ALTER EXTENSION vector UPDATE TO '0.8.0'" to load this file. \quit
CREATE FUNCTION hnsw_intvec_support(internal) RETURNS internal
AS 'MODULE_PATHNAME' LANGUAGE C;
CREATE TYPE intvec;
CREATE FUNCTION intvec_in(cstring, oid, integer) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_out(intvec) RETURNS cstring
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_typmod_in(cstring[]) RETURNS integer
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_recv(internal, oid, integer) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_send(intvec) RETURNS bytea
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE TYPE intvec (
INPUT = intvec_in,
OUTPUT = intvec_out,
TYPMOD_IN = intvec_typmod_in,
RECEIVE = intvec_recv,
SEND = intvec_send,
STORAGE = external
);
CREATE FUNCTION l2_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_l2_distance' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION inner_product(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_inner_product' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION cosine_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_cosine_distance' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION l1_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_l1_distance' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION vector_dims(intvec) RETURNS integer
AS 'MODULE_PATHNAME', 'intvec_vector_dims' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION l2_norm(intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_l2_norm' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_lt(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_le(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_eq(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_ne(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_ge(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_gt(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_cmp(intvec, intvec) RETURNS int4
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_l2_squared_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_negative_inner_product(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec(intvec, integer, boolean) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION array_to_intvec(integer[], integer, boolean) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_to_integer(intvec, integer, boolean) RETURNS integer[]
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE CAST (intvec AS intvec)
WITH FUNCTION intvec(intvec, integer, boolean) AS IMPLICIT;
CREATE CAST (intvec AS integer[])
WITH FUNCTION intvec_to_integer(intvec, integer, boolean) AS ASSIGNMENT;
CREATE CAST (integer[] AS intvec)
WITH FUNCTION array_to_intvec(integer[], integer, boolean) AS ASSIGNMENT;
CREATE OPERATOR <-> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = l2_distance,
COMMUTATOR = '<->'
);
CREATE OPERATOR <#> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_negative_inner_product,
COMMUTATOR = '<#>'
);
CREATE OPERATOR <=> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = cosine_distance,
COMMUTATOR = '<=>'
);
CREATE OPERATOR <+> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = l1_distance,
COMMUTATOR = '<+>'
);
CREATE OPERATOR < (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_lt,
COMMUTATOR = > , NEGATOR = >= ,
RESTRICT = scalarltsel, JOIN = scalarltjoinsel
);
CREATE OPERATOR <= (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_le,
COMMUTATOR = >= , NEGATOR = > ,
RESTRICT = scalarlesel, JOIN = scalarlejoinsel
);
CREATE OPERATOR = (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_eq,
COMMUTATOR = = , NEGATOR = <> ,
RESTRICT = eqsel, JOIN = eqjoinsel
);
CREATE OPERATOR <> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_ne,
COMMUTATOR = <> , NEGATOR = = ,
RESTRICT = eqsel, JOIN = eqjoinsel
);
CREATE OPERATOR >= (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_ge,
COMMUTATOR = <= , NEGATOR = < ,
RESTRICT = scalargesel, JOIN = scalargejoinsel
);
CREATE OPERATOR > (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_gt,
COMMUTATOR = < , NEGATOR = <= ,
RESTRICT = scalargtsel, JOIN = scalargtjoinsel
);
CREATE OPERATOR CLASS intvec_ops
DEFAULT FOR TYPE intvec USING btree AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 intvec_cmp(intvec, intvec);
CREATE OPERATOR CLASS intvec_l2_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <-> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 intvec_l2_squared_distance(intvec, intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE OPERATOR CLASS intvec_ip_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <#> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 intvec_negative_inner_product(intvec, intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE OPERATOR CLASS intvec_cosine_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <=> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 cosine_distance(intvec, intvec),
FUNCTION 2 l2_norm(intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE OPERATOR CLASS intvec_l1_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <+> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 l1_distance(intvec, intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE FUNCTION array_to_sparsevec(integer[], integer, boolean) RETURNS sparsevec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;

View File

@@ -272,9 +272,6 @@ CREATE FUNCTION ivfflat_bit_support(internal) RETURNS internal
CREATE FUNCTION hnsw_halfvec_support(internal) RETURNS internal
AS 'MODULE_PATHNAME' LANGUAGE C;
CREATE FUNCTION hnsw_intvec_support(internal) RETURNS internal
AS 'MODULE_PATHNAME' LANGUAGE C;
CREATE FUNCTION hnsw_bit_support(internal) RETURNS internal
AS 'MODULE_PATHNAME' LANGUAGE C;
@@ -650,199 +647,6 @@ CREATE OPERATOR CLASS halfvec_l1_ops
FUNCTION 1 l1_distance(halfvec, halfvec),
FUNCTION 3 hnsw_halfvec_support(internal);
-- intvec type
CREATE TYPE intvec;
CREATE FUNCTION intvec_in(cstring, oid, integer) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_out(intvec) RETURNS cstring
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_typmod_in(cstring[]) RETURNS integer
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_recv(internal, oid, integer) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_send(intvec) RETURNS bytea
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE TYPE intvec (
INPUT = intvec_in,
OUTPUT = intvec_out,
TYPMOD_IN = intvec_typmod_in,
RECEIVE = intvec_recv,
SEND = intvec_send,
STORAGE = external
);
-- intvec functions
CREATE FUNCTION l2_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_l2_distance' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION inner_product(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_inner_product' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION cosine_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_cosine_distance' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION l1_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_l1_distance' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION vector_dims(intvec) RETURNS integer
AS 'MODULE_PATHNAME', 'intvec_vector_dims' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION l2_norm(intvec) RETURNS float8
AS 'MODULE_PATHNAME', 'intvec_l2_norm' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
-- intvec private functions
CREATE FUNCTION intvec_lt(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_le(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_eq(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_ne(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_ge(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_gt(intvec, intvec) RETURNS bool
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_cmp(intvec, intvec) RETURNS int4
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_l2_squared_distance(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_negative_inner_product(intvec, intvec) RETURNS float8
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
-- intvec cast functions
CREATE FUNCTION intvec(intvec, integer, boolean) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION array_to_intvec(integer[], integer, boolean) RETURNS intvec
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
CREATE FUNCTION intvec_to_integer(intvec, integer, boolean) RETURNS integer[]
AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
-- intvec casts
CREATE CAST (intvec AS intvec)
WITH FUNCTION intvec(intvec, integer, boolean) AS IMPLICIT;
CREATE CAST (intvec AS integer[])
WITH FUNCTION intvec_to_integer(intvec, integer, boolean) AS ASSIGNMENT;
CREATE CAST (integer[] AS intvec)
WITH FUNCTION array_to_intvec(integer[], integer, boolean) AS ASSIGNMENT;
-- intvec operators
CREATE OPERATOR <-> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = l2_distance,
COMMUTATOR = '<->'
);
CREATE OPERATOR <#> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_negative_inner_product,
COMMUTATOR = '<#>'
);
CREATE OPERATOR <=> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = cosine_distance,
COMMUTATOR = '<=>'
);
CREATE OPERATOR <+> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = l1_distance,
COMMUTATOR = '<+>'
);
CREATE OPERATOR < (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_lt,
COMMUTATOR = > , NEGATOR = >= ,
RESTRICT = scalarltsel, JOIN = scalarltjoinsel
);
CREATE OPERATOR <= (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_le,
COMMUTATOR = >= , NEGATOR = > ,
RESTRICT = scalarlesel, JOIN = scalarlejoinsel
);
CREATE OPERATOR = (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_eq,
COMMUTATOR = = , NEGATOR = <> ,
RESTRICT = eqsel, JOIN = eqjoinsel
);
CREATE OPERATOR <> (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_ne,
COMMUTATOR = <> , NEGATOR = = ,
RESTRICT = eqsel, JOIN = eqjoinsel
);
CREATE OPERATOR >= (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_ge,
COMMUTATOR = <= , NEGATOR = < ,
RESTRICT = scalargesel, JOIN = scalargejoinsel
);
CREATE OPERATOR > (
LEFTARG = intvec, RIGHTARG = intvec, PROCEDURE = intvec_gt,
COMMUTATOR = < , NEGATOR = <= ,
RESTRICT = scalargtsel, JOIN = scalargtjoinsel
);
-- intvec opclasses
CREATE OPERATOR CLASS intvec_ops
DEFAULT FOR TYPE intvec USING btree AS
OPERATOR 1 < ,
OPERATOR 2 <= ,
OPERATOR 3 = ,
OPERATOR 4 >= ,
OPERATOR 5 > ,
FUNCTION 1 intvec_cmp(intvec, intvec);
CREATE OPERATOR CLASS intvec_l2_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <-> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 intvec_l2_squared_distance(intvec, intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE OPERATOR CLASS intvec_ip_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <#> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 intvec_negative_inner_product(intvec, intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE OPERATOR CLASS intvec_cosine_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <=> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 cosine_distance(intvec, intvec),
FUNCTION 2 l2_norm(intvec),
FUNCTION 3 hnsw_intvec_support(internal);
CREATE OPERATOR CLASS intvec_l1_ops
FOR TYPE intvec USING hnsw AS
OPERATOR 1 <+> (intvec, intvec) FOR ORDER BY float_ops,
FUNCTION 1 l1_distance(intvec, intvec),
FUNCTION 3 hnsw_intvec_support(internal);
-- bit functions
CREATE FUNCTION hamming_distance(bit, bit) RETURNS float8

View File

@@ -159,6 +159,24 @@ CheckStateArray(ArrayType *statearray, const char *caller)
return (float8 *) ARR_DATA_PTR(statearray);
}
#if PG_VERSION_NUM < 120003
static pg_noinline void
float_overflow_error(void)
{
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("value out of range: overflow")));
}
static pg_noinline void
float_underflow_error(void)
{
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("value out of range: underflow")));
}
#endif
/*
* Convert textual representation to internal representation
*/

View File

@@ -12,22 +12,12 @@
#include "utils/float.h"
#include "utils/guc.h"
#include "utils/selfuncs.h"
#include "utils/spccache.h"
#if PG_VERSION_NUM < 150000
#define MarkGUCPrefixReserved(x) EmitWarningsOnPlaceholders(x)
#endif
static const struct config_enum_entry hnsw_iterative_search_options[] = {
{"off", HNSW_ITERATIVE_SEARCH_OFF, false},
{"relaxed_order", HNSW_ITERATIVE_SEARCH_RELAXED, false},
{"strict_order", HNSW_ITERATIVE_SEARCH_STRICT, false},
{NULL, 0, false}
};
int hnsw_ef_search;
int hnsw_max_search_tuples;
int hnsw_iterative_search;
int hnsw_lock_tranche_id;
static relopt_kind hnsw_relopt_kind;
@@ -78,15 +68,6 @@ HnswInit(void)
"Valid range is 1..1000.", &hnsw_ef_search,
HNSW_DEFAULT_EF_SEARCH, HNSW_MIN_EF_SEARCH, HNSW_MAX_EF_SEARCH, PGC_USERSET, 0, NULL, NULL, NULL);
DefineCustomEnumVariable("hnsw.iterative_search", "Sets the iterative search mode",
NULL, &hnsw_iterative_search,
HNSW_ITERATIVE_SEARCH_OFF, hnsw_iterative_search_options, PGC_USERSET, 0, NULL, NULL, NULL);
/* This is approximate and does not apply to the initial scan */
DefineCustomIntVariable("hnsw.max_search_tuples", "Sets the max number of candidates to visit for iterative search",
"-1 means no limit", &hnsw_max_search_tuples,
-1, -1, INT_MAX, PGC_USERSET, 0, NULL, NULL, NULL);
MarkGUCPrefixReserved("hnsw");
}
@@ -118,9 +99,7 @@ hnswcostestimate(PlannerInfo *root, IndexPath *path, double loop_count,
{
GenericCosts costs;
int m;
double ratio;
double startupPages;
double spc_seq_page_cost;
int entryLevel;
Relation index;
/* Never use index without order */
@@ -136,71 +115,21 @@ hnswcostestimate(PlannerInfo *root, IndexPath *path, double loop_count,
MemSet(&costs, 0, sizeof(costs));
genericcostestimate(root, path, loop_count, &costs);
index = index_open(path->indexinfo->indexoid, NoLock);
HnswGetMetaPageInfo(index, &m, NULL);
index_close(index, NoLock);
/*
* HNSW cost estimation follows a formula that accounts for the total
* number of tuples indexed combined with the parameters that most
* influence the duration of the index scan, namely: m - the number of
* tuples that are scanned in each step of the HNSW graph traversal
* ef_search - which influences the total number of steps taken at layer 0
*
* The source of the vector data can impact how many steps it takes to
* converge on the set of vectors to return to the executor. Currently, we
* use a hardcoded scaling factor (HNSWScanScalingFactor) to help
* influence that, but this could later become a configurable parameter
* based on the cost estimations.
*
* The tuple estimator formula is below:
*
* numIndexTuples = entryLevel * m + layer0TuplesMax * layer0Selectivity
*
* "entryLevel * m" represents the floor of tuples we need to scan to get
* to layer 0 (L0).
*
* "layer0TuplesMax" is the estimated total number of tuples we'd scan at
* L0 if we weren't discarding already visited tuples as part of the scan.
*
* "layer0Selectivity" estimates the percentage of tuples that are scanned
* at L0, accounting for previously visited tuples, multiplied by the
* "scalingFactor" (currently hardcoded).
*/
if (path->indexinfo->tuples > 0)
{
double scalingFactor = 0.55;
int entryLevel = (int) (log(path->indexinfo->tuples) * HnswGetMl(m));
int layer0TuplesMax = HnswGetLayerM(m, 0) * hnsw_ef_search;
double layer0Selectivity = scalingFactor * log(path->indexinfo->tuples) / (log(m) * (1 + log(hnsw_ef_search)));
/* Approximate entry level */
entryLevel = (int) -log(1.0 / path->indexinfo->tuples) * HnswGetMl(m);
ratio = (entryLevel * m + layer0TuplesMax * layer0Selectivity) / path->indexinfo->tuples;
/* TODO Improve estimate of visited tuples (currently underestimates) */
/* Account for number of tuples (or entry level), m, and ef_search */
costs.numIndexTuples = (entryLevel + 2) * m;
if (ratio > 1)
ratio = 1;
}
else
ratio = 1;
genericcostestimate(root, path, loop_count, &costs);
get_tablespace_page_costs(path->indexinfo->reltablespace, NULL, &spc_seq_page_cost);
/* Startup cost is cost before returning the first row */
costs.indexStartupCost = costs.indexTotalCost * ratio;
/* Adjust cost if needed since TOAST not included in seq scan cost */
startupPages = costs.numIndexPages * ratio;
if (startupPages > path->indexinfo->rel->pages && ratio < 0.5)
{
/* Change all page cost from random to sequential */
costs.indexStartupCost -= startupPages * (costs.spc_random_page_cost - spc_seq_page_cost);
/* Remove cost of extra pages */
costs.indexStartupCost -= (startupPages - path->indexinfo->rel->pages) * spc_seq_page_cost;
}
*indexStartupCost = costs.indexStartupCost;
/* Use total cost since most work happens before first tuple is returned */
*indexStartupCost = costs.indexTotalCost;
*indexTotalCost = costs.indexTotalCost;
*indexSelectivity = costs.indexSelectivity;
*indexCorrelation = costs.indexCorrelation;

View File

@@ -88,9 +88,6 @@
/* Ensure fits on page and in uint8 */
#define HnswGetMaxLevel(m) Min(((BLCKSZ - MAXALIGN(SizeOfPageHeaderData) - MAXALIGN(sizeof(HnswPageOpaqueData)) - offsetof(HnswNeighborTupleData, indextids) - sizeof(ItemIdData)) / (sizeof(ItemPointerData)) / (m)) - 2, 255)
#define HnswGetSearchCandidate(membername, ptr) pairingheap_container(HnswSearchCandidate, membername, ptr)
#define HnswGetSearchCandidateConst(membername, ptr) pairingheap_const_container(HnswSearchCandidate, membername, ptr)
#define HnswGetValue(base, element) PointerGetDatum(HnswPtrAccess(base, (element)->value))
#if PG_VERSION_NUM < 140005
@@ -109,17 +106,8 @@
/* Variables */
extern int hnsw_ef_search;
extern int hnsw_iterative_search;
extern int hnsw_max_search_tuples;
extern int hnsw_lock_tranche_id;
typedef enum HnswIterativeSearchMode
{
HNSW_ITERATIVE_SEARCH_OFF,
HNSW_ITERATIVE_SEARCH_RELAXED,
HNSW_ITERATIVE_SEARCH_STRICT
} HnswIterativeSearchMode;
typedef struct HnswElementData HnswElementData;
typedef struct HnswNeighborArray HnswNeighborArray;
@@ -141,7 +129,6 @@ struct HnswElementData
uint8 heaptidsLength;
uint8 level;
uint8 deleted;
uint8 version;
uint32 hash;
HnswNeighborsPtr neighbors;
BlockNumber blkno;
@@ -173,7 +160,7 @@ typedef struct HnswSearchCandidate
pairingheap_node c_node;
pairingheap_node w_node;
HnswElementPtr element;
double distance;
float distance;
} HnswSearchCandidate;
/* HNSW index options */
@@ -198,8 +185,8 @@ typedef struct HnswGraph
/* Allocations state */
LWLock allocatorLock;
Size memoryUsed;
Size memoryTotal;
long memoryUsed;
long memoryTotal;
/* Flushed state */
LWLock flushLock;
@@ -250,18 +237,6 @@ typedef struct HnswTypeInfo
void (*checkValue) (Pointer v);
} HnswTypeInfo;
typedef struct HnswSupport
{
FmgrInfo *procinfo;
FmgrInfo *normprocinfo;
Oid collation;
} HnswSupport;
typedef struct HnswQuery
{
Datum value;
} HnswQuery;
typedef struct HnswBuildState
{
/* Info */
@@ -281,7 +256,9 @@ typedef struct HnswBuildState
double reltuples;
/* Support functions */
HnswSupport support;
FmgrInfo *procinfo;
FmgrInfo *normprocinfo;
Oid collation;
/* Variables */
HnswGraph graphData;
@@ -329,10 +306,10 @@ typedef struct HnswElementTupleData
uint8 type;
uint8 level;
uint8 deleted;
uint8 version;
uint8 unused;
ItemPointerData heaptids[HNSW_HEAPTIDS];
ItemPointerData neighbortid;
uint16 unused;
uint16 unused2;
Vector data;
} HnswElementTupleData;
@@ -341,41 +318,24 @@ typedef HnswElementTupleData * HnswElementTuple;
typedef struct HnswNeighborTupleData
{
uint8 type;
uint8 version;
uint8 unused;
uint16 count;
ItemPointerData indextids[FLEXIBLE_ARRAY_MEMBER];
} HnswNeighborTupleData;
typedef HnswNeighborTupleData * HnswNeighborTuple;
typedef union
{
struct pointerhash_hash *pointers;
struct offsethash_hash *offsets;
struct tidhash_hash *tids;
} visited_hash;
typedef union
{
HnswElement element;
ItemPointerData indextid;
} HnswUnvisited;
typedef struct HnswScanOpaqueData
{
const HnswTypeInfo *typeInfo;
bool first;
List *w;
visited_hash v;
pairingheap *discarded;
HnswQuery q;
int m;
int64 tuples;
double previousDistance;
MemoryContext tmpCtx;
/* Support functions */
HnswSupport support;
FmgrInfo *procinfo;
FmgrInfo *normprocinfo;
Oid collation;
} HnswScanOpaqueData;
typedef HnswScanOpaqueData * HnswScanOpaque;
@@ -393,7 +353,8 @@ typedef struct HnswVacuumState
int efConstruction;
/* Support functions */
HnswSupport support;
FmgrInfo *procinfo;
Oid collation;
/* Variables */
struct tidhash_hash *deleted;
@@ -409,33 +370,30 @@ typedef struct HnswVacuumState
int HnswGetM(Relation index);
int HnswGetEfConstruction(Relation index);
FmgrInfo *HnswOptionalProcInfo(Relation index, uint16 procnum);
void HnswInitSupport(HnswSupport * support, Relation index);
Datum HnswNormValue(const HnswTypeInfo * typeInfo, Oid collation, Datum value);
bool HnswCheckNorm(HnswSupport * support, Datum value);
bool HnswCheckNorm(FmgrInfo *procinfo, Oid collation, Datum value);
Buffer HnswNewBuffer(Relation index, ForkNumber forkNum);
void HnswInitPage(Buffer buf, Page page);
void HnswInit(void);
List *HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation index, HnswSupport * support, int m, bool inserting, HnswElement skipElement, visited_hash * v, pairingheap **discarded, bool initVisited, int64 *tuples);
List *HnswSearchLayer(char *base, Datum q, List *ep, int ef, int lc, Relation index, FmgrInfo *procinfo, Oid collation, int m, bool inserting, HnswElement skipElement);
HnswElement HnswGetEntryPoint(Relation index);
void HnswGetMetaPageInfo(Relation index, int *m, HnswElement * entryPoint);
void *HnswAlloc(HnswAllocator * allocator, Size size);
HnswElement HnswInitElement(char *base, ItemPointer tid, int m, double ml, int maxLevel, HnswAllocator * alloc);
HnswElement HnswInitElementFromBlock(BlockNumber blkno, OffsetNumber offno);
void HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, HnswSupport * support, int m, int efConstruction, bool existing);
HnswSearchCandidate *HnswEntryCandidate(char *base, HnswElement em, HnswQuery * q, Relation rel, HnswSupport * support, bool loadVec);
void HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, FmgrInfo *procinfo, Oid collation, int m, int efConstruction, bool existing);
HnswSearchCandidate *HnswEntryCandidate(char *base, HnswElement em, Datum q, Relation rel, FmgrInfo *procinfo, Oid collation, bool loadVec);
void HnswUpdateMetaPage(Relation index, int updateEntry, HnswElement entryPoint, BlockNumber insertPage, ForkNumber forkNum, bool building);
void HnswSetNeighborTuple(char *base, HnswNeighborTuple ntup, HnswElement e, int m);
void HnswAddHeapTid(HnswElement element, ItemPointer heaptid);
HnswNeighborArray *HnswInitNeighborArray(int lm, HnswAllocator * allocator);
void HnswInitNeighbors(char *base, HnswElement element, int m, HnswAllocator * alloc);
bool HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPointer heaptid, bool building);
void HnswUpdateNeighborsOnDisk(Relation index, HnswSupport * support, HnswElement e, int m, bool checkExisting, bool building);
bool HnswInsertTupleOnDisk(Relation index, Datum value, Datum *values, bool *isnull, ItemPointer heap_tid, bool building);
void HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, HnswElement e, int m, bool checkExisting, bool building);
void HnswLoadElementFromTuple(HnswElement element, HnswElementTuple etup, bool loadHeaptids, bool loadVec);
void HnswLoadElement(HnswElement element, double *distance, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec, double *maxDistance);
bool HnswFormIndexValue(Datum *out, Datum *values, bool *isnull, const HnswTypeInfo * typeInfo, HnswSupport * support);
void HnswLoadElement(HnswElement element, float *distance, Datum *q, Relation index, FmgrInfo *procinfo, Oid collation, bool loadVec, float *maxDistance);
void HnswSetElementTuple(char *base, HnswElementTuple etup, HnswElement element);
void HnswUpdateConnection(char *base, HnswNeighborArray * neighbors, HnswElement newElement, float distance, int lm, int *updateIdx, Relation index, HnswSupport * support);
bool HnswLoadNeighborTids(HnswElement element, ItemPointerData *indextids, Relation index, int m, int lm, int lc);
void HnswUpdateConnection(char *base, HnswElement element, HnswCandidate * hc, int lm, int lc, int *updateIdx, Relation index, FmgrInfo *procinfo, Oid collation);
void HnswLoadNeighbors(HnswElement element, Relation index, int m);
void HnswInitLockTranche(void);
const HnswTypeInfo *HnswGetTypeInfo(Relation index);
PGDLLEXPORT void HnswParallelBuildMain(dsm_segment *seg, shm_toc *toc);

View File

@@ -366,7 +366,7 @@ AddElementInMemory(char *base, HnswGraph * graph, HnswElement element)
* Update neighbors
*/
static void
UpdateNeighborsInMemory(char *base, HnswSupport * support, HnswElement e, int m)
UpdateNeighborsInMemory(char *base, FmgrInfo *procinfo, Oid collation, HnswElement e, int m)
{
for (int lc = e->level; lc >= 0; lc--)
{
@@ -388,7 +388,7 @@ UpdateNeighborsInMemory(char *base, HnswSupport * support, HnswElement e, int m)
Assert(neighborElement);
LWLockAcquire(&neighborElement->lock, LW_EXCLUSIVE);
HnswUpdateConnection(base, HnswGetNeighbors(base, neighborElement, lc), e, hc->distance, lm, NULL, NULL, support);
HnswUpdateConnection(base, e, hc, lm, lc, NULL, NULL, procinfo, collation);
LWLockRelease(&neighborElement->lock);
}
}
@@ -398,7 +398,7 @@ UpdateNeighborsInMemory(char *base, HnswSupport * support, HnswElement e, int m)
* Update graph in memory
*/
static void
UpdateGraphInMemory(HnswSupport * support, HnswElement element, int m, int efConstruction, HnswElement entryPoint, HnswBuildState * buildstate)
UpdateGraphInMemory(FmgrInfo *procinfo, Oid collation, HnswElement element, int m, int efConstruction, HnswElement entryPoint, HnswBuildState * buildstate)
{
HnswGraph *graph = buildstate->graph;
char *base = buildstate->hnswarea;
@@ -411,7 +411,7 @@ UpdateGraphInMemory(HnswSupport * support, HnswElement element, int m, int efCon
AddElementInMemory(base, graph, element);
/* Update neighbors */
UpdateNeighborsInMemory(base, support, element, m);
UpdateNeighborsInMemory(base, procinfo, collation, element, m);
/* Update entry point if needed (already have lock) */
if (entryPoint == NULL || element->level > entryPoint->level)
@@ -424,8 +424,9 @@ UpdateGraphInMemory(HnswSupport * support, HnswElement element, int m, int efCon
static void
InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
{
FmgrInfo *procinfo = buildstate->procinfo;
Oid collation = buildstate->collation;
HnswGraph *graph = buildstate->graph;
HnswSupport *support = &buildstate->support;
HnswElement entryPoint;
LWLock *entryLock = &graph->entryLock;
LWLock *entryWaitLock = &graph->entryWaitLock;
@@ -457,10 +458,10 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
}
/* Find neighbors for element */
HnswFindElementNeighbors(base, element, entryPoint, NULL, support, m, efConstruction, false);
HnswFindElementNeighbors(base, element, entryPoint, NULL, procinfo, collation, m, efConstruction, false);
/* Update graph in memory */
UpdateGraphInMemory(support, element, m, efConstruction, entryPoint, buildstate);
UpdateGraphInMemory(procinfo, collation, element, m, efConstruction, entryPoint, buildstate);
/* Release entry lock */
LWLockRelease(entryLock);
@@ -472,19 +473,30 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
static bool
InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heaptid, HnswBuildState * buildstate)
{
const HnswTypeInfo *typeInfo = buildstate->typeInfo;
HnswGraph *graph = buildstate->graph;
HnswElement element;
HnswAllocator *allocator = &buildstate->allocator;
HnswSupport *support = &buildstate->support;
Size valueSize;
Pointer valuePtr;
LWLock *flushLock = &graph->flushLock;
char *base = buildstate->hnswarea;
Datum value;
/* Form index value */
if (!HnswFormIndexValue(&value, values, isnull, buildstate->typeInfo, support))
return false;
/* Detoast once for all calls */
Datum value = PointerGetDatum(PG_DETOAST_DATUM(values[0]));
/* Check value */
if (typeInfo->checkValue != NULL)
typeInfo->checkValue(DatumGetPointer(value));
/* Normalize if needed */
if (buildstate->normprocinfo != NULL)
{
if (!HnswCheckNorm(buildstate->normprocinfo, buildstate->collation, value))
return false;
value = HnswNormValue(typeInfo, buildstate->collation, value);
}
/* Get datum size */
valueSize = VARSIZE_ANY(DatumGetPointer(value));
@@ -497,7 +509,7 @@ InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heaptid, Hn
{
LWLockRelease(flushLock);
return HnswInsertTupleOnDisk(index, support, value, heaptid, true);
return HnswInsertTupleOnDisk(index, value, values, isnull, heaptid, true);
}
/*
@@ -529,7 +541,7 @@ InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heaptid, Hn
LWLockRelease(flushLock);
return HnswInsertTupleOnDisk(index, support, value, heaptid, true);
return HnswInsertTupleOnDisk(index, value, values, isnull, heaptid, true);
}
/* Ok, we can proceed to allocate the element */
@@ -595,7 +607,7 @@ BuildCallback(Relation index, ItemPointer tid, Datum *values,
* Initialize the graph
*/
static void
InitGraph(HnswGraph * graph, char *base, Size memoryTotal)
InitGraph(HnswGraph * graph, char *base, long memoryTotal)
{
/* Initialize the lock tranche if needed */
HnswInitLockTranche();
@@ -692,9 +704,11 @@ InitBuildState(HnswBuildState * buildstate, Relation heap, Relation index, Index
buildstate->indtuples = 0;
/* Get support functions */
HnswInitSupport(&buildstate->support, index);
buildstate->procinfo = index_getprocinfo(index, 1, HNSW_DISTANCE_PROC);
buildstate->normprocinfo = HnswOptionalProcInfo(index, HNSW_NORM_PROC);
buildstate->collation = index->rd_indcollation[0];
InitGraph(&buildstate->graphData, NULL, (Size) maintenance_work_mem * 1024L);
InitGraph(&buildstate->graphData, NULL, maintenance_work_mem * 1024L);
buildstate->graph = &buildstate->graphData;
buildstate->ml = HnswGetMl(buildstate->m);
buildstate->maxLevel = HnswGetMaxLevel(buildstate->m);

View File

@@ -36,7 +36,7 @@ GetInsertPage(Relation index)
* Check for a free offset
*/
static bool
HnswFreeOffset(Relation index, Buffer buf, Page page, HnswElement element, Size etupSize, Size ntupSize, Buffer *nbuf, Page *npage, OffsetNumber *freeOffno, OffsetNumber *freeNeighborOffno, BlockNumber *newInsertPage, uint8 *tupleVersion)
HnswFreeOffset(Relation index, Buffer buf, Page page, HnswElement element, Size etupSize, Size ntupSize, Buffer *nbuf, Page *npage, OffsetNumber *freeOffno, OffsetNumber *freeNeighborOffno, BlockNumber *newInsertPage)
{
OffsetNumber offno;
OffsetNumber maxoffno = PageGetMaxOffsetNumber(page);
@@ -98,7 +98,6 @@ HnswFreeOffset(Relation index, Buffer buf, Page page, HnswElement element, Size
{
*freeOffno = offno;
*freeNeighborOffno = neighborOffno;
*tupleVersion = etup->version;
return true;
}
else if (*nbuf != buf)
@@ -154,7 +153,6 @@ AddElementOnDisk(Relation index, HnswElement e, int m, BlockNumber insertPage, B
OffsetNumber freeOffno = InvalidOffsetNumber;
OffsetNumber freeNeighborOffno = InvalidOffsetNumber;
BlockNumber newInsertPage = InvalidBlockNumber;
uint8 tupleVersion;
char *base = NULL;
/* Calculate sizes */
@@ -204,7 +202,7 @@ AddElementOnDisk(Relation index, HnswElement e, int m, BlockNumber insertPage, B
}
/* Next, try space from a deleted element */
if (HnswFreeOffset(index, buf, page, e, etupSize, ntupSize, &nbuf, &npage, &freeOffno, &freeNeighborOffno, &newInsertPage, &tupleVersion))
if (HnswFreeOffset(index, buf, page, e, etupSize, ntupSize, &nbuf, &npage, &freeOffno, &freeNeighborOffno, &newInsertPage))
{
if (nbuf != buf)
{
@@ -214,10 +212,6 @@ AddElementOnDisk(Relation index, HnswElement e, int m, BlockNumber insertPage, B
npage = GenericXLogRegisterBuffer(state, nbuf, 0);
}
/* Set tuple version */
etup->version = tupleVersion;
ntup->version = tupleVersion;
break;
}
@@ -340,107 +334,6 @@ AddElementOnDisk(Relation index, HnswElement e, int m, BlockNumber insertPage, B
*updatedInsertPage = newInsertPage;
}
/*
* Load neighbors
*/
static HnswNeighborArray *
HnswLoadNeighbors(HnswElement element, Relation index, int m, int lm, int lc)
{
char *base = NULL;
HnswNeighborArray *neighbors = HnswInitNeighborArray(lm, NULL);
ItemPointerData indextids[HNSW_MAX_M * 2];
if (!HnswLoadNeighborTids(element, indextids, index, m, lm, lc))
return neighbors;
for (int i = 0; i < lm; i++)
{
ItemPointer indextid = &indextids[i];
HnswElement e;
HnswCandidate *hc;
if (!ItemPointerIsValid(indextid))
break;
e = HnswInitElementFromBlock(ItemPointerGetBlockNumber(indextid), ItemPointerGetOffsetNumber(indextid));
hc = &neighbors->items[neighbors->length++];
HnswPtrStore(base, hc->element, e);
}
return neighbors;
}
/*
* Load elements for insert
*/
static void
LoadElementsForInsert(HnswNeighborArray * neighbors, HnswQuery * q, int *idx, Relation index, HnswSupport * support)
{
char *base = NULL;
for (int i = 0; i < neighbors->length; i++)
{
HnswCandidate *hc = &neighbors->items[i];
HnswElement element = HnswPtrAccess(base, hc->element);
double distance;
HnswLoadElement(element, &distance, q, index, support, true, NULL);
hc->distance = distance;
/* Prune element if being deleted */
if (element->heaptidsLength == 0)
{
*idx = i;
break;
}
}
}
/*
* Get update index
*/
static int
GetUpdateIndex(HnswElement element, HnswElement newElement, float distance, int m, int lm, int lc, Relation index, HnswSupport * support, MemoryContext updateCtx)
{
char *base = NULL;
int idx = -1;
HnswNeighborArray *neighbors;
MemoryContext oldCtx = MemoryContextSwitchTo(updateCtx);
/*
* Get latest neighbors since they may have changed. Do not lock yet since
* selecting neighbors can take time. Could use optimistic locking to
* retry if another update occurs before getting exclusive lock.
*/
neighbors = HnswLoadNeighbors(element, index, m, lm, lc);
/*
* Could improve performance for vacuuming by checking neighbors against
* list of elements being deleted to find index. It's important to exclude
* already deleted elements for this since they can be replaced at any
* time.
*/
if (neighbors->length < lm)
idx = -2;
else
{
HnswQuery q;
q.value = HnswGetValue(base, element);
LoadElementsForInsert(neighbors, &q, &idx, index, support);
if (idx == -1)
HnswUpdateConnection(base, neighbors, newElement, distance, lm, &idx, index, support);
}
MemoryContextSwitchTo(oldCtx);
MemoryContextReset(updateCtx);
return idx;
}
/*
* Check if connection already exists
*/
@@ -461,94 +354,14 @@ ConnectionExists(HnswElement e, HnswNeighborTuple ntup, int startIdx, int lm)
return false;
}
/*
* Update neighbor
*/
static void
UpdateNeighborOnDisk(HnswElement element, HnswElement newElement, int idx, int m, int lm, int lc, Relation index, bool checkExisting, bool building)
{
Buffer buf;
Page page;
GenericXLogState *state;
HnswNeighborTuple ntup;
int startIdx;
OffsetNumber offno = element->neighborOffno;
/* Register page */
buf = ReadBuffer(index, element->neighborPage);
LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
if (building)
{
state = NULL;
page = BufferGetPage(buf);
}
else
{
state = GenericXLogStart(index);
page = GenericXLogRegisterBuffer(state, buf, 0);
}
/* Get tuple */
ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, offno));
/* Calculate index for update */
startIdx = (element->level - lc) * m;
/* Check for existing connection */
if (checkExisting && ConnectionExists(newElement, ntup, startIdx, lm))
idx = -1;
else if (idx == -2)
{
/* Find free offset if still exists */
/* TODO Retry updating connections if not */
for (int j = 0; j < lm; j++)
{
if (!ItemPointerIsValid(&ntup->indextids[startIdx + j]))
{
idx = startIdx + j;
break;
}
}
}
else
idx += startIdx;
/* Make robust to issues */
if (idx >= 0 && idx < ntup->count)
{
ItemPointer indextid = &ntup->indextids[idx];
/* Update neighbor on the buffer */
ItemPointerSet(indextid, newElement->blkno, newElement->offno);
/* Commit */
if (building)
MarkBufferDirty(buf);
else
GenericXLogFinish(state);
}
else if (!building)
GenericXLogAbort(state);
UnlockReleaseBuffer(buf);
}
/*
* Update neighbors
*/
void
HnswUpdateNeighborsOnDisk(Relation index, HnswSupport * support, HnswElement e, int m, bool checkExisting, bool building)
HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, HnswElement e, int m, bool checkExisting, bool building)
{
char *base = NULL;
/* Use separate memory context to improve performance for larger vectors */
MemoryContext updateCtx = GenerationContextCreate(CurrentMemoryContext,
"Hnsw insert update context",
#if PG_VERSION_NUM >= 150000
128 * 1024, 128 * 1024,
#endif
128 * 1024);
for (int lc = e->level; lc >= 0; lc--)
{
int lm = HnswGetLayerM(m, lc);
@@ -557,20 +370,96 @@ HnswUpdateNeighborsOnDisk(Relation index, HnswSupport * support, HnswElement e,
for (int i = 0; i < neighbors->length; i++)
{
HnswCandidate *hc = &neighbors->items[i];
Buffer buf;
Page page;
GenericXLogState *state;
HnswNeighborTuple ntup;
int idx = -1;
int startIdx;
HnswElement neighborElement = HnswPtrAccess(base, hc->element);
int idx;
OffsetNumber offno = neighborElement->neighborOffno;
idx = GetUpdateIndex(neighborElement, e, hc->distance, m, lm, lc, index, support, updateCtx);
/*
* Get latest neighbors since they may have changed. Do not lock
* yet since selecting neighbors can take time. Could use
* optimistic locking to retry if another update occurs before
* getting exclusive lock.
*/
HnswLoadNeighbors(neighborElement, index, m);
/*
* Could improve performance for vacuuming by checking neighbors
* against list of elements being deleted to find index. It's
* important to exclude already deleted elements for this since
* they can be replaced at any time.
*/
/* Select neighbors */
HnswUpdateConnection(NULL, e, hc, lm, lc, &idx, index, procinfo, collation);
/* New element was not selected as a neighbor */
if (idx == -1)
continue;
UpdateNeighborOnDisk(neighborElement, e, idx, m, lm, lc, index, checkExisting, building);
/* Register page */
buf = ReadBuffer(index, neighborElement->neighborPage);
LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
if (building)
{
state = NULL;
page = BufferGetPage(buf);
}
else
{
state = GenericXLogStart(index);
page = GenericXLogRegisterBuffer(state, buf, 0);
}
/* Get tuple */
ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, offno));
/* Calculate index for update */
startIdx = (neighborElement->level - lc) * m;
/* Check for existing connection */
if (checkExisting && ConnectionExists(e, ntup, startIdx, lm))
idx = -1;
else if (idx == -2)
{
/* Find free offset if still exists */
/* TODO Retry updating connections if not */
for (int j = 0; j < lm; j++)
{
if (!ItemPointerIsValid(&ntup->indextids[startIdx + j]))
{
idx = startIdx + j;
break;
}
}
}
else
idx += startIdx;
/* Make robust to issues */
if (idx >= 0 && idx < ntup->count)
{
ItemPointer indextid = &ntup->indextids[idx];
/* Update neighbor on the buffer */
ItemPointerSet(indextid, e->blkno, e->offno);
/* Commit */
if (building)
MarkBufferDirty(buf);
else
GenericXLogFinish(state);
}
else if (!building)
GenericXLogAbort(state);
UnlockReleaseBuffer(buf);
}
}
MemoryContextDelete(updateCtx);
}
/*
@@ -660,7 +549,7 @@ FindDuplicateOnDisk(Relation index, HnswElement element, bool building)
* Update graph on disk
*/
static void
UpdateGraphOnDisk(Relation index, HnswSupport * support, HnswElement element, int m, int efConstruction, HnswElement entryPoint, bool building)
UpdateGraphOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, HnswElement element, int m, int efConstruction, HnswElement entryPoint, bool building)
{
BlockNumber newInsertPage = InvalidBlockNumber;
@@ -676,7 +565,7 @@ UpdateGraphOnDisk(Relation index, HnswSupport * support, HnswElement element, in
HnswUpdateMetaPage(index, 0, NULL, newInsertPage, MAIN_FORKNUM, building);
/* Update neighbors */
HnswUpdateNeighborsOnDisk(index, support, element, m, false, building);
HnswUpdateNeighborsOnDisk(index, procinfo, collation, element, m, false, building);
/* Update entry point if needed */
if (entryPoint == NULL || element->level > entryPoint->level)
@@ -687,12 +576,14 @@ UpdateGraphOnDisk(Relation index, HnswSupport * support, HnswElement element, in
* Insert a tuple into the index
*/
bool
HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPointer heaptid, bool building)
HnswInsertTupleOnDisk(Relation index, Datum value, Datum *values, bool *isnull, ItemPointer heap_tid, bool building)
{
HnswElement entryPoint;
HnswElement element;
int m;
int efConstruction = HnswGetEfConstruction(index);
FmgrInfo *procinfo = index_getprocinfo(index, 1, HNSW_DISTANCE_PROC);
Oid collation = index->rd_indcollation[0];
LOCKMODE lockmode = ShareLock;
char *base = NULL;
@@ -707,7 +598,7 @@ HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPo
HnswGetMetaPageInfo(index, &m, &entryPoint);
/* Create an element */
element = HnswInitElement(base, heaptid, m, HnswGetMl(m), HnswGetMaxLevel(m), NULL);
element = HnswInitElement(base, heap_tid, m, HnswGetMl(m), HnswGetMaxLevel(m), NULL);
HnswPtrStore(base, element->value, DatumGetPointer(value));
/* Prevent concurrent inserts when likely updating entry point */
@@ -725,10 +616,10 @@ HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPo
}
/* Find neighbors for element */
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, false);
HnswFindElementNeighbors(base, element, entryPoint, index, procinfo, collation, m, efConstruction, false);
/* Update graph on disk */
UpdateGraphOnDisk(index, support, element, m, efConstruction, entryPoint, building);
UpdateGraphOnDisk(index, procinfo, collation, element, m, efConstruction, entryPoint, building);
/* Release lock */
UnlockPage(index, HNSW_UPDATE_LOCK, lockmode);
@@ -740,19 +631,31 @@ HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPo
* Insert a tuple into the index
*/
static void
HnswInsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heaptid)
HnswInsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heap_tid)
{
Datum value;
const HnswTypeInfo *typeInfo = HnswGetTypeInfo(index);
HnswSupport support;
FmgrInfo *normprocinfo;
Oid collation = index->rd_indcollation[0];
HnswInitSupport(&support, index);
/* Detoast once for all calls */
value = PointerGetDatum(PG_DETOAST_DATUM(values[0]));
/* Form index value */
if (!HnswFormIndexValue(&value, values, isnull, typeInfo, &support))
return;
/* Check value */
if (typeInfo->checkValue != NULL)
typeInfo->checkValue(DatumGetPointer(value));
HnswInsertTupleOnDisk(index, &support, value, heaptid, false);
/* Normalize if needed */
normprocinfo = HnswOptionalProcInfo(index, HNSW_NORM_PROC);
if (normprocinfo != NULL)
{
if (!HnswCheckNorm(normprocinfo, collation, value))
return;
value = HnswNormValue(typeInfo, collation, value);
}
HnswInsertTupleOnDisk(index, value, values, isnull, heap_tid, false);
}
/*

View File

@@ -5,74 +5,39 @@
#include "pgstat.h"
#include "storage/bufmgr.h"
#include "storage/lmgr.h"
#include "utils/float.h"
#include "utils/memutils.h"
/*
* Algorithm 5 from paper
*/
static List *
GetScanItems(IndexScanDesc scan, Datum value)
GetScanItems(IndexScanDesc scan, Datum q)
{
HnswScanOpaque so = (HnswScanOpaque) scan->opaque;
Relation index = scan->indexRelation;
HnswSupport *support = &so->support;
FmgrInfo *procinfo = so->procinfo;
Oid collation = so->collation;
List *ep;
List *w;
int m;
HnswElement entryPoint;
char *base = NULL;
HnswQuery *q = &so->q;
/* Get m and entry point */
HnswGetMetaPageInfo(index, &m, &entryPoint);
q->value = value;
so->m = m;
if (entryPoint == NULL)
return NIL;
ep = list_make1(HnswEntryCandidate(base, entryPoint, q, index, support, false));
ep = list_make1(HnswEntryCandidate(base, entryPoint, q, index, procinfo, collation, false));
for (int lc = entryPoint->level; lc >= 1; lc--)
{
w = HnswSearchLayer(base, q, ep, 1, lc, index, support, m, false, NULL, NULL, NULL, true, NULL);
w = HnswSearchLayer(base, q, ep, 1, lc, index, procinfo, collation, m, false, NULL);
ep = w;
}
return HnswSearchLayer(base, q, ep, hnsw_ef_search, 0, index, support, m, false, NULL, &so->v, hnsw_iterative_search != HNSW_ITERATIVE_SEARCH_OFF ? &so->discarded : NULL, true, &so->tuples);
}
/*
* Resume scan at ground level with discarded candidates
*/
static List *
ResumeScanItems(IndexScanDesc scan)
{
HnswScanOpaque so = (HnswScanOpaque) scan->opaque;
Relation index = scan->indexRelation;
List *ep = NIL;
char *base = NULL;
int batch_size = hnsw_ef_search;
if (pairingheap_is_empty(so->discarded))
return NIL;
/* Get next batch of candidates */
for (int i = 0; i < batch_size; i++)
{
HnswSearchCandidate *sc;
if (pairingheap_is_empty(so->discarded))
break;
sc = HnswGetSearchCandidate(w_node, pairingheap_remove_first(so->discarded));
ep = lappend(ep, sc);
}
return HnswSearchLayer(base, &so->q, ep, batch_size, 0, index, &so->support, so->m, false, NULL, &so->v, &so->discarded, false, &so->tuples);
return HnswSearchLayer(base, q, ep, hnsw_ef_search, 0, index, procinfo, collation, m, false, NULL);
}
/*
@@ -95,24 +60,13 @@ GetScanValue(IndexScanDesc scan)
Assert(!VARATT_IS_EXTENDED(DatumGetPointer(value)));
/* Normalize if needed */
if (so->support.normprocinfo != NULL)
value = HnswNormValue(so->typeInfo, so->support.collation, value);
if (so->normprocinfo != NULL)
value = HnswNormValue(so->typeInfo, so->collation, value);
}
return value;
}
#if defined(HNSW_MEMORY)
/*
* Show memory usage
*/
static void
ShowMemoryUsage(HnswScanOpaque so)
{
elog(INFO, "memory: %zu KB, tuples: " INT64_FORMAT, MemoryContextMemAllocated(so->tmpCtx, false) / 1024, so->tuples);
}
#endif
/*
* Prepare for an index scan
*/
@@ -127,19 +81,14 @@ hnswbeginscan(Relation index, int nkeys, int norderbys)
so = (HnswScanOpaque) palloc(sizeof(HnswScanOpaqueData));
so->typeInfo = HnswGetTypeInfo(index);
so->first = true;
so->v.tids = NULL;
so->discarded = NULL;
/*
* Use a lower max allocation size than default to allow scanning more
* tuples for iterative search before exceeding work_mem
*/
so->tmpCtx = AllocSetContextCreate(CurrentMemoryContext,
"Hnsw scan temporary context",
0, 8 * 1024, 512 * 1024);
ALLOCSET_DEFAULT_SIZES);
/* Set support functions */
HnswInitSupport(&so->support, index);
so->procinfo = index_getprocinfo(index, 1, HNSW_DISTANCE_PROC);
so->normprocinfo = HnswOptionalProcInfo(index, HNSW_NORM_PROC);
so->collation = index->rd_indcollation[0];
scan->opaque = so;
@@ -154,15 +103,7 @@ hnswrescan(IndexScanDesc scan, ScanKey keys, int nkeys, ScanKey orderbys, int no
{
HnswScanOpaque so = (HnswScanOpaque) scan->opaque;
if (so->v.tids != NULL)
tidhash_reset(so->v.tids);
if (so->discarded != NULL)
pairingheap_reset(so->discarded);
so->first = true;
so->tuples = 0;
so->previousDistance = -get_float8_infinity();
MemoryContextReset(so->tmpCtx);
if (keys && scan->numberOfKeys > 0)
@@ -220,104 +161,26 @@ hnswgettuple(IndexScanDesc scan, ScanDirection dir)
so->first = false;
#if defined(HNSW_MEMORY)
ShowMemoryUsage(so);
elog(INFO, "memory: %zu KB", MemoryContextMemAllocated(so->tmpCtx, false) / 1024);
#endif
}
for (;;)
while (list_length(so->w) > 0)
{
char *base = NULL;
HnswSearchCandidate *sc;
HnswElement element;
HnswSearchCandidate *hc = llast(so->w);
HnswElement element = HnswPtrAccess(base, hc->element);
ItemPointer heaptid;
if (list_length(so->w) == 0)
{
if (hnsw_iterative_search == HNSW_ITERATIVE_SEARCH_OFF)
break;
/* Empty index */
if (so->discarded == NULL)
break;
/* Reached max number of tuples */
if (hnsw_max_search_tuples != -1 && so->tuples >= hnsw_max_search_tuples)
{
if (pairingheap_is_empty(so->discarded))
break;
/* Return remaining tuples */
so->w = lappend(so->w, HnswGetSearchCandidate(w_node, pairingheap_remove_first(so->discarded)));
}
/* Prevent scans from consuming too much memory */
else if (MemoryContextMemAllocated(so->tmpCtx, false) > (Size) work_mem * 1024L)
{
if (pairingheap_is_empty(so->discarded))
{
ereport(DEBUG1,
(errmsg("hnsw index scan exceeded work_mem after " INT64_FORMAT " tuples", so->tuples),
errhint("Increase work_mem to scan more tuples.")));
break;
}
/* Return remaining tuples */
so->w = lappend(so->w, HnswGetSearchCandidate(w_node, pairingheap_remove_first(so->discarded)));
}
else
{
/*
* Locking ensures when neighbors are read, the elements they
* reference will not be deleted (and replaced) during the
* iteration.
*
* Elements loaded into memory on previous iterations may have
* been deleted (and replaced), so when reading neighbors, the
* element version must be checked.
*/
LockPage(scan->indexRelation, HNSW_SCAN_LOCK, ShareLock);
so->w = ResumeScanItems(scan);
UnlockPage(scan->indexRelation, HNSW_SCAN_LOCK, ShareLock);
#if defined(HNSW_MEMORY)
ShowMemoryUsage(so);
#endif
}
if (list_length(so->w) == 0)
break;
}
sc = llast(so->w);
element = HnswPtrAccess(base, sc->element);
/* Move to next element if no valid heap TIDs */
if (element->heaptidsLength == 0)
{
so->w = list_delete_last(so->w);
/* Mark memory as free for next iteration */
if (hnsw_iterative_search != HNSW_ITERATIVE_SEARCH_OFF)
{
pfree(element);
pfree(sc);
}
continue;
}
heaptid = &element->heaptids[--element->heaptidsLength];
if (hnsw_iterative_search == HNSW_ITERATIVE_SEARCH_STRICT)
{
if (sc->distance < so->previousDistance)
continue;
so->previousDistance = sc->distance;
}
MemoryContextSwitchTo(oldCtx);
scan->xs_heaptid = *heaptid;

View File

@@ -1,6 +1,5 @@
#include "postgres.h"
#include <float.h>
#include <math.h>
#include "access/generic_xlog.h"
@@ -101,6 +100,19 @@ hash_offset(Size offset)
#define SH_DEFINE
#include "lib/simplehash.h"
typedef union
{
pointerhash_hash *pointers;
offsethash_hash *offsets;
tidhash_hash *tids;
} visited_hash;
typedef union
{
HnswElement element;
ItemPointerData indextid;
} HnswUnvisited;
/*
* Get the max number of connections in an upper layer for each element in the index
*/
@@ -141,26 +153,12 @@ HnswOptionalProcInfo(Relation index, uint16 procnum)
return index_getprocinfo(index, 1, procnum);
}
/*
* Init support functions
*/
void
HnswInitSupport(HnswSupport * support, Relation index)
{
support->procinfo = index_getprocinfo(index, 1, HNSW_DISTANCE_PROC);
support->collation = index->rd_indcollation[0];
support->normprocinfo = HnswOptionalProcInfo(index, HNSW_NORM_PROC);
}
/*
* Normalize value
*/
Datum
HnswNormValue(const HnswTypeInfo * typeInfo, Oid collation, Datum value)
{
if (!typeInfo->normalize)
return value;
return DirectFunctionCall1Coll(typeInfo->normalize, collation, value);
}
@@ -168,9 +166,9 @@ HnswNormValue(const HnswTypeInfo * typeInfo, Oid collation, Datum value)
* Check if non-zero norm
*/
bool
HnswCheckNorm(HnswSupport * support, Datum value)
HnswCheckNorm(FmgrInfo *procinfo, Oid collation, Datum value)
{
return DatumGetFloat8(FunctionCall1Coll(support->normprocinfo, support->collation, value)) > 0;
return DatumGetFloat8(FunctionCall1Coll(procinfo, collation, value)) > 0;
}
/*
@@ -199,7 +197,7 @@ HnswInitPage(Buffer buf, Page page)
/*
* Allocate a neighbor array
*/
HnswNeighborArray *
static HnswNeighborArray *
HnswInitNeighborArray(int lm, HnswAllocator * allocator)
{
HnswNeighborArray *a = HnswAlloc(allocator, HNSW_NEIGHBOR_ARRAY_SIZE(lm));
@@ -255,8 +253,6 @@ HnswInitElement(char *base, ItemPointer heaptid, int m, double ml, int maxLevel,
element->level = level;
element->deleted = 0;
/* Start at one to make it easier to find issues */
element->version = 1;
HnswInitNeighbors(base, element, m, allocator);
@@ -398,33 +394,6 @@ HnswUpdateMetaPage(Relation index, int updateEntry, HnswElement entryPoint, Bloc
UnlockReleaseBuffer(buf);
}
/*
* Form index value
*/
bool
HnswFormIndexValue(Datum *out, Datum *values, bool *isnull, const HnswTypeInfo * typeInfo, HnswSupport * support)
{
/* Detoast once for all calls */
Datum value = PointerGetDatum(PG_DETOAST_DATUM(values[0]));
/* Check value */
if (typeInfo->checkValue != NULL)
typeInfo->checkValue(DatumGetPointer(value));
/* Normalize if needed */
if (support->normprocinfo != NULL)
{
if (!HnswCheckNorm(support, value))
return false;
value = HnswNormValue(typeInfo, support->collation, value);
}
*out = value;
return true;
}
/*
* Set element tuple, except for neighbor info
*/
@@ -436,7 +405,6 @@ HnswSetElementTuple(char *base, HnswElementTuple etup, HnswElement element)
etup->type = HNSW_ELEMENT_TUPLE_TYPE;
etup->level = element->level;
etup->deleted = 0;
etup->version = element->version;
for (int i = 0; i < HNSW_HEAPTIDS; i++)
{
if (i < element->heaptidsLength)
@@ -479,7 +447,69 @@ HnswSetNeighborTuple(char *base, HnswNeighborTuple ntup, HnswElement e, int m)
}
ntup->count = idx;
ntup->version = e->version;
}
/*
* Load neighbors from page
*/
static void
LoadNeighborsFromPage(HnswElement element, Relation index, Page page, int m)
{
char *base = NULL;
HnswNeighborTuple ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, element->neighborOffno));
int neighborCount = (element->level + 2) * m;
Assert(HnswIsNeighborTuple(ntup));
HnswInitNeighbors(base, element, m, NULL);
/* Ensure expected neighbors */
if (ntup->count != neighborCount)
return;
for (int i = 0; i < neighborCount; i++)
{
HnswElement e;
int level;
HnswCandidate *hc;
ItemPointer indextid;
HnswNeighborArray *neighbors;
indextid = &ntup->indextids[i];
if (!ItemPointerIsValid(indextid))
continue;
e = HnswInitElementFromBlock(ItemPointerGetBlockNumber(indextid), ItemPointerGetOffsetNumber(indextid));
/* Calculate level based on offset */
level = element->level - i / m;
if (level < 0)
level = 0;
neighbors = HnswGetNeighbors(base, element, level);
hc = &neighbors->items[neighbors->length++];
HnswPtrStore(base, hc->element, e);
}
}
/*
* Load neighbors
*/
void
HnswLoadNeighbors(HnswElement element, Relation index, int m)
{
Buffer buf;
Page page;
buf = ReadBuffer(index, element->neighborPage);
LockBuffer(buf, BUFFER_LOCK_SHARE);
page = BufferGetPage(buf);
LoadNeighborsFromPage(element, index, page, m);
UnlockReleaseBuffer(buf);
}
/*
@@ -490,7 +520,6 @@ HnswLoadElementFromTuple(HnswElement element, HnswElementTuple etup, bool loadHe
{
element->level = etup->level;
element->deleted = etup->deleted;
element->version = etup->version;
element->neighborPage = ItemPointerGetBlockNumber(&etup->neighbortid);
element->neighborOffno = ItemPointerGetOffsetNumber(&etup->neighbortid);
element->heaptidsLength = 0;
@@ -516,20 +545,11 @@ HnswLoadElementFromTuple(HnswElement element, HnswElementTuple etup, bool loadHe
}
}
/*
* Calculate the distance between values
*/
static inline double
HnswGetDistance(Datum a, Datum b, HnswSupport * support)
{
return DatumGetFloat8(FunctionCall2Coll(support->procinfo, support->collation, a, b));
}
/*
* Load an element and optionally get its distance from q
*/
static void
HnswLoadElementImpl(BlockNumber blkno, OffsetNumber offno, double *distance, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec, double *maxDistance, HnswElement * element)
HnswLoadElementImpl(BlockNumber blkno, OffsetNumber offno, float *distance, Datum *q, Relation index, FmgrInfo *procinfo, Oid collation, bool loadVec, float *maxDistance, HnswElement * element)
{
Buffer buf;
Page page;
@@ -547,17 +567,10 @@ HnswLoadElementImpl(BlockNumber blkno, OffsetNumber offno, double *distance, Hns
/* Calculate distance */
if (distance != NULL)
{
if (DatumGetPointer(q->value) == NULL)
if (DatumGetPointer(*q) == NULL)
*distance = 0;
else
{
*distance = HnswGetDistance(q->value, PointerGetDatum(&etup->data), support);
/* Needed for intvec cosine distance */
/* TODO Improve */
if (isnan(*distance))
*distance = DBL_MAX;
}
*distance = (float) DatumGetFloat8(FunctionCall2Coll(procinfo, collation, *q, PointerGetDatum(&etup->data)));
}
/* Load element */
@@ -576,52 +589,41 @@ HnswLoadElementImpl(BlockNumber blkno, OffsetNumber offno, double *distance, Hns
* Load an element and optionally get its distance from q
*/
void
HnswLoadElement(HnswElement element, double *distance, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec, double *maxDistance)
HnswLoadElement(HnswElement element, float *distance, Datum *q, Relation index, FmgrInfo *procinfo, Oid collation, bool loadVec, float *maxDistance)
{
HnswLoadElementImpl(element->blkno, element->offno, distance, q, index, support, loadVec, maxDistance, &element);
HnswLoadElementImpl(element->blkno, element->offno, distance, q, index, procinfo, collation, loadVec, maxDistance, &element);
}
/*
* Get the distance for an element
*/
static double
GetElementDistance(char *base, HnswElement element, HnswQuery * q, HnswSupport * support)
static float
GetElementDistance(char *base, HnswElement element, Datum q, FmgrInfo *procinfo, Oid collation)
{
Datum value = HnswGetValue(base, element);
return HnswGetDistance(q->value, value, support);
}
/*
* Allocate a search candidate
*/
static HnswSearchCandidate *
HnswInitSearchCandidate(char *base, HnswElement element, double distance)
{
HnswSearchCandidate *sc = palloc(sizeof(HnswSearchCandidate));
HnswPtrStore(base, sc->element, element);
sc->distance = distance;
return sc;
return DatumGetFloat8(FunctionCall2Coll(procinfo, collation, q, value));
}
/*
* Create a candidate for the entry point
*/
HnswSearchCandidate *
HnswEntryCandidate(char *base, HnswElement entryPoint, HnswQuery * q, Relation index, HnswSupport * support, bool loadVec)
HnswEntryCandidate(char *base, HnswElement entryPoint, Datum q, Relation index, FmgrInfo *procinfo, Oid collation, bool loadVec)
{
bool inMemory = index == NULL;
double distance;
HnswSearchCandidate *hc = palloc(sizeof(HnswSearchCandidate));
if (inMemory)
distance = GetElementDistance(base, entryPoint, q, support);
HnswPtrStore(base, hc->element, entryPoint);
if (index == NULL)
hc->distance = GetElementDistance(base, entryPoint, q, procinfo, collation);
else
HnswLoadElement(entryPoint, &distance, q, index, support, loadVec, NULL);
return HnswInitSearchCandidate(base, entryPoint, distance);
HnswLoadElement(entryPoint, &hc->distance, &q, index, procinfo, collation, loadVec, NULL);
return hc;
}
#define HnswGetSearchCandidate(membername, ptr) pairingheap_container(HnswSearchCandidate, membername, ptr)
#define HnswGetSearchCandidateConst(membername, ptr) pairingheap_const_container(HnswSearchCandidate, membername, ptr)
/*
* Compare candidate distances
*/
@@ -637,21 +639,6 @@ CompareNearestCandidates(const pairingheap_node *a, const pairingheap_node *b, v
return 0;
}
/*
* Compare discarded candidate distances
*/
static int
CompareNearestDiscardedCandidates(const pairingheap_node *a, const pairingheap_node *b, void *arg)
{
if (HnswGetSearchCandidateConst(w_node, a)->distance < HnswGetSearchCandidateConst(w_node, b)->distance)
return 1;
if (HnswGetSearchCandidateConst(w_node, a)->distance > HnswGetSearchCandidateConst(w_node, b)->distance)
return -1;
return 0;
}
/*
* Compare candidate distances
*/
@@ -671,9 +658,9 @@ CompareFurthestCandidates(const pairingheap_node *a, const pairingheap_node *b,
* Init visited
*/
static inline void
InitVisited(char *base, visited_hash * v, bool inMemory, int ef, int m)
InitVisited(char *base, visited_hash * v, Relation index, int ef, int m)
{
if (!inMemory)
if (index != NULL)
v->tids = tidhash_create(CurrentMemoryContext, ef * m * 2, NULL);
else if (base != NULL)
v->offsets = offsethash_create(CurrentMemoryContext, ef * m * 2, NULL);
@@ -685,9 +672,9 @@ InitVisited(char *base, visited_hash * v, bool inMemory, int ef, int m)
* Add to visited
*/
static inline void
AddToVisited(char *base, visited_hash * v, HnswElementPtr elementPtr, bool inMemory, bool *found)
AddToVisited(char *base, visited_hash * v, HnswElementPtr elementPtr, Relation index, bool *found)
{
if (!inMemory)
if (index != NULL)
{
HnswElement element = HnswPtrAccess(base, elementPtr);
ItemPointerData indextid;
@@ -748,60 +735,38 @@ HnswLoadUnvisitedFromMemory(char *base, HnswElement element, HnswUnvisited * unv
HnswCandidate *hc = &localNeighborhood->items[i];
bool found;
AddToVisited(base, v, hc->element, true, &found);
AddToVisited(base, v, hc->element, NULL, &found);
if (!found)
unvisited[(*unvisitedLength)++].element = HnswPtrAccess(base, hc->element);
}
}
/*
* Load neighbor index TIDs
*/
bool
HnswLoadNeighborTids(HnswElement element, ItemPointerData *indextids, Relation index, int m, int lm, int lc)
{
Buffer buf;
Page page;
HnswNeighborTuple ntup;
int start;
buf = ReadBuffer(index, element->neighborPage);
LockBuffer(buf, BUFFER_LOCK_SHARE);
page = BufferGetPage(buf);
ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, element->neighborOffno));
/*
* Ensure the neighbor tuple has not been deleted or replaced between
* index scan iterations
*/
if (ntup->version != element->version || ntup->count != (element->level + 2) * m)
{
UnlockReleaseBuffer(buf);
return false;
}
/* Copy to minimize lock time */
start = (element->level - lc) * m;
memcpy(indextids, ntup->indextids + start, lm * sizeof(ItemPointerData));
UnlockReleaseBuffer(buf);
return true;
}
/*
* Load unvisited neighbors from disk
*/
static void
HnswLoadUnvisitedFromDisk(HnswElement element, HnswUnvisited * unvisited, int *unvisitedLength, visited_hash * v, Relation index, int m, int lm, int lc)
{
Buffer buf;
Page page;
HnswNeighborTuple ntup;
int start;
ItemPointerData indextids[HNSW_MAX_M * 2];
*unvisitedLength = 0;
buf = ReadBuffer(index, element->neighborPage);
LockBuffer(buf, BUFFER_LOCK_SHARE);
page = BufferGetPage(buf);
if (!HnswLoadNeighborTids(element, indextids, index, m, lm, lc))
return;
ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, element->neighborOffno));
start = (element->level - lc) * m;
/* Copy to minimize lock time */
memcpy(&indextids, ntup->indextids + start, lm * sizeof(ItemPointerData));
UnlockReleaseBuffer(buf);
*unvisitedLength = 0;
for (int i = 0; i < lm; i++)
{
@@ -822,37 +787,24 @@ HnswLoadUnvisitedFromDisk(HnswElement element, HnswUnvisited * unvisited, int *u
* Algorithm 2 from paper
*/
List *
HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation index, HnswSupport * support, int m, bool inserting, HnswElement skipElement, visited_hash * v, pairingheap **discarded, bool initVisited, int64 *tuples)
HnswSearchLayer(char *base, Datum q, List *ep, int ef, int lc, Relation index, FmgrInfo *procinfo, Oid collation, int m, bool inserting, HnswElement skipElement)
{
List *w = NIL;
pairingheap *C = pairingheap_allocate(CompareNearestCandidates, NULL);
pairingheap *W = pairingheap_allocate(CompareFurthestCandidates, NULL);
int wlen = 0;
visited_hash vh;
visited_hash v;
ListCell *lc2;
HnswNeighborArray *localNeighborhood = NULL;
Size neighborhoodSize = 0;
int lm = HnswGetLayerM(m, lc);
HnswUnvisited *unvisited = palloc(lm * sizeof(HnswUnvisited));
int unvisitedLength;
bool inMemory = index == NULL;
if (v == NULL)
{
v = &vh;
initVisited = true;
}
if (initVisited)
{
InitVisited(base, v, inMemory, ef, m);
if (discarded != NULL)
*discarded = pairingheap_allocate(CompareNearestDiscardedCandidates, NULL);
}
InitVisited(base, &v, index, ef, m);
/* Create local memory for neighborhood if needed */
if (inMemory)
if (index == NULL)
{
neighborhoodSize = HNSW_NEIGHBOR_ARRAY_SIZE(lm);
localNeighborhood = palloc(neighborhoodSize);
@@ -861,26 +813,20 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
/* Add entry points to v, C, and W */
foreach(lc2, ep)
{
HnswSearchCandidate *sc = (HnswSearchCandidate *) lfirst(lc2);
HnswSearchCandidate *hc = (HnswSearchCandidate *) lfirst(lc2);
bool found;
if (initVisited)
{
AddToVisited(base, v, sc->element, inMemory, &found);
AddToVisited(base, &v, hc->element, index, &found);
if (tuples != NULL)
(*tuples)++;
}
pairingheap_add(C, &sc->c_node);
pairingheap_add(W, &sc->w_node);
pairingheap_add(C, &hc->c_node);
pairingheap_add(W, &hc->w_node);
/*
* Do not count elements being deleted towards ef when vacuuming. It
* would be ideal to do this for inserts as well, but this could
* affect insert performance.
*/
if (CountElement(skipElement, HnswPtrAccess(base, sc->element)))
if (CountElement(skipElement, HnswPtrAccess(base, hc->element)))
wlen++;
}
@@ -895,27 +841,24 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
cElement = HnswPtrAccess(base, c->element);
if (inMemory)
HnswLoadUnvisitedFromMemory(base, cElement, unvisited, &unvisitedLength, v, lc, localNeighborhood, neighborhoodSize);
if (index == NULL)
HnswLoadUnvisitedFromMemory(base, cElement, unvisited, &unvisitedLength, &v, lc, localNeighborhood, neighborhoodSize);
else
HnswLoadUnvisitedFromDisk(cElement, unvisited, &unvisitedLength, v, index, m, lm, lc);
if (tuples != NULL)
(*tuples) += unvisitedLength;
HnswLoadUnvisitedFromDisk(cElement, unvisited, &unvisitedLength, &v, index, m, lm, lc);
for (int i = 0; i < unvisitedLength; i++)
{
HnswElement eElement;
HnswSearchCandidate *e;
double eDistance;
float eDistance;
bool alwaysAdd = wlen < ef;
f = HnswGetSearchCandidate(w_node, pairingheap_first(W));
if (inMemory)
if (index == NULL)
{
eElement = unvisited[i].element;
eDistance = GetElementDistance(base, eElement, q, support);
eDistance = GetElementDistance(base, eElement, q, procinfo, collation);
}
else
{
@@ -925,30 +868,25 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
/* Avoid any allocations if not adding */
eElement = NULL;
HnswLoadElementImpl(blkno, offno, &eDistance, q, index, support, inserting, alwaysAdd || discarded != NULL ? NULL : &f->distance, &eElement);
HnswLoadElementImpl(blkno, offno, &eDistance, &q, index, procinfo, collation, inserting, alwaysAdd ? NULL : &f->distance, &eElement);
if (eElement == NULL)
continue;
}
if (eElement == NULL || !(eDistance < f->distance || alwaysAdd))
{
if (discarded != NULL)
{
/* Create a new candidate */
e = HnswInitSearchCandidate(base, eElement, eDistance);
pairingheap_add(*discarded, &e->w_node);
}
if (!(eDistance < f->distance || alwaysAdd))
continue;
}
Assert(!eElement->deleted);
/* Make robust to issues */
if (eElement->level < lc)
continue;
/* Create a new candidate */
e = HnswInitSearchCandidate(base, eElement, eDistance);
e = palloc(sizeof(HnswSearchCandidate));
HnswPtrStore(base, e->element, eElement);
e->distance = eDistance;
pairingheap_add(C, &e->c_node);
pairingheap_add(W, &e->w_node);
@@ -963,12 +901,7 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
/* No need to decrement wlen */
if (wlen > ef)
{
HnswSearchCandidate *d = HnswGetSearchCandidate(w_node, pairingheap_remove_first(W));
if (discarded != NULL)
pairingheap_add(*discarded, &d->w_node);
}
pairingheap_remove_first(W);
}
}
}
@@ -976,9 +909,9 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
/* Add each element of W to w */
while (!pairingheap_is_empty(W))
{
HnswSearchCandidate *sc = HnswGetSearchCandidate(w_node, pairingheap_remove_first(W));
HnswSearchCandidate *hc = HnswGetSearchCandidate(w_node, pairingheap_remove_first(W));
w = lappend(w, sc);
w = lappend(w, hc);
}
return w;
@@ -1032,22 +965,32 @@ CompareCandidateDistancesOffset(const ListCell *a, const ListCell *b)
return 0;
}
/*
* Calculate the distance between elements
*/
static float
HnswGetDistance(char *base, HnswElement a, HnswElement b, FmgrInfo *procinfo, Oid collation)
{
Datum aValue = HnswGetValue(base, a);
Datum bValue = HnswGetValue(base, b);
return DatumGetFloat8(FunctionCall2Coll(procinfo, collation, aValue, bValue));
}
/*
* Check if an element is closer to q than any element from R
*/
static bool
CheckElementCloser(char *base, HnswCandidate * e, List *r, HnswSupport * support)
CheckElementCloser(char *base, HnswCandidate * e, List *r, FmgrInfo *procinfo, Oid collation)
{
HnswElement eElement = HnswPtrAccess(base, e->element);
Datum eValue = HnswGetValue(base, eElement);
ListCell *lc2;
foreach(lc2, r)
{
HnswCandidate *ri = lfirst(lc2);
HnswElement riElement = HnswPtrAccess(base, ri->element);
Datum riValue = HnswGetValue(base, riElement);
float distance = HnswGetDistance(eValue, riValue, support);
float distance = HnswGetDistance(base, eElement, riElement, procinfo, collation);
if (distance <= e->distance)
return false;
@@ -1060,14 +1003,15 @@ CheckElementCloser(char *base, HnswCandidate * e, List *r, HnswSupport * support
* Algorithm 4 from paper
*/
static List *
SelectNeighbors(char *base, List *c, int lm, HnswSupport * support, bool *closerSet, HnswCandidate * newCandidate, HnswCandidate * *pruned, bool sortCandidates)
SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid collation, HnswElement e2, HnswCandidate * newCandidate, HnswCandidate * *pruned, bool sortCandidates)
{
List *r = NIL;
List *w = list_copy(c);
HnswCandidate **wd;
int wdlen = 0;
int wdoff = 0;
bool mustCalculate = !(*closerSet);
HnswNeighborArray *neighbors = HnswGetNeighbors(base, e2, lc);
bool mustCalculate = !neighbors->closerSet;
List *added = NIL;
bool removedAny = false;
@@ -1094,7 +1038,7 @@ SelectNeighbors(char *base, List *c, int lm, HnswSupport * support, bool *closer
/* Use previous state of r and wd to skip work when possible */
if (mustCalculate)
e->closer = CheckElementCloser(base, e, r, support);
e->closer = CheckElementCloser(base, e, r, procinfo, collation);
else if (list_length(added) > 0)
{
/* Keep Valgrind happy for in-memory, parallel builds */
@@ -1107,7 +1051,7 @@ SelectNeighbors(char *base, List *c, int lm, HnswSupport * support, bool *closer
*/
if (e->closer)
{
e->closer = CheckElementCloser(base, e, added, support);
e->closer = CheckElementCloser(base, e, added, procinfo, collation);
if (!e->closer)
removedAny = true;
@@ -1120,7 +1064,7 @@ SelectNeighbors(char *base, List *c, int lm, HnswSupport * support, bool *closer
*/
if (removedAny)
{
e->closer = CheckElementCloser(base, e, r, support);
e->closer = CheckElementCloser(base, e, r, procinfo, collation);
if (e->closer)
added = lappend(added, e);
}
@@ -1128,7 +1072,7 @@ SelectNeighbors(char *base, List *c, int lm, HnswSupport * support, bool *closer
}
else if (e == newCandidate)
{
e->closer = CheckElementCloser(base, e, r, support);
e->closer = CheckElementCloser(base, e, r, procinfo, collation);
if (e->closer)
added = lappend(added, e);
}
@@ -1144,7 +1088,7 @@ SelectNeighbors(char *base, List *c, int lm, HnswSupport * support, bool *closer
}
/* Cached value can only be used in future if sorted deterministically */
*closerSet = sortCandidates;
neighbors->closerSet = sortCandidates;
/* Keep pruned connections */
while (wdoff < wdlen && list_length(r) < lm)
@@ -1179,16 +1123,18 @@ AddConnections(char *base, HnswElement element, List *neighbors, int lc)
* Update connections
*/
void
HnswUpdateConnection(char *base, HnswNeighborArray * neighbors, HnswElement newElement, float distance, int lm, int *updateIdx, Relation index, HnswSupport * support)
HnswUpdateConnection(char *base, HnswElement element, HnswCandidate * hc, int lm, int lc, int *updateIdx, Relation index, FmgrInfo *procinfo, Oid collation)
{
HnswCandidate newHc;
HnswElement hce = HnswPtrAccess(base, hc->element);
HnswNeighborArray *currentNeighbors = HnswGetNeighbors(base, hce, lc);
HnswCandidate hc2;
HnswPtrStore(base, newHc.element, newElement);
newHc.distance = distance;
HnswPtrStore(base, hc2.element, element);
hc2.distance = hc->distance;
if (neighbors->length < lm)
if (currentNeighbors->length < lm)
{
neighbors->items[neighbors->length++] = newHc;
currentNeighbors->items[currentNeighbors->length++] = hc2;
/* Track update */
if (updateIdx != NULL)
@@ -1197,26 +1143,54 @@ HnswUpdateConnection(char *base, HnswNeighborArray * neighbors, HnswElement newE
else
{
/* Shrink connections */
List *c = NIL;
HnswCandidate *pruned = NULL;
/* Add candidates */
for (int i = 0; i < neighbors->length; i++)
c = lappend(c, &neighbors->items[i]);
c = lappend(c, &newHc);
/* Load elements on insert */
if (index != NULL)
{
Datum q = HnswGetValue(base, hce);
SelectNeighbors(base, c, lm, support, &neighbors->closerSet, &newHc, &pruned, true);
for (int i = 0; i < currentNeighbors->length; i++)
{
HnswCandidate *hc3 = &currentNeighbors->items[i];
HnswElement hc3Element = HnswPtrAccess(base, hc3->element);
if (HnswPtrIsNull(base, hc3Element->value))
HnswLoadElement(hc3Element, &hc3->distance, &q, index, procinfo, collation, true, NULL);
else
hc3->distance = GetElementDistance(base, hc3Element, q, procinfo, collation);
/* Prune element if being deleted */
if (hc3Element->heaptidsLength == 0)
{
pruned = &currentNeighbors->items[i];
break;
}
}
}
/* Should not happen */
if (pruned == NULL)
return;
{
List *c = NIL;
/* Add candidates */
for (int i = 0; i < currentNeighbors->length; i++)
c = lappend(c, &currentNeighbors->items[i]);
c = lappend(c, &hc2);
SelectNeighbors(base, c, lm, lc, procinfo, collation, hce, &hc2, &pruned, true);
/* Should not happen */
if (pruned == NULL)
return;
}
/* Find and replace the pruned element */
for (int i = 0; i < neighbors->length; i++)
for (int i = 0; i < currentNeighbors->length; i++)
{
if (HnswPtrEqual(base, neighbors->items[i].element, pruned->element))
if (HnswPtrEqual(base, currentNeighbors->items[i].element, pruned->element))
{
neighbors->items[i] = newHc;
currentNeighbors->items[i] = hc2;
/* Track update */
if (updateIdx != NULL)
@@ -1276,20 +1250,17 @@ PrecomputeHash(char *base, HnswElement element)
* Algorithm 1 from paper
*/
void
HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, HnswSupport * support, int m, int efConstruction, bool existing)
HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint, Relation index, FmgrInfo *procinfo, Oid collation, int m, int efConstruction, bool existing)
{
List *ep;
List *w;
int level = element->level;
int entryLevel;
HnswQuery q;
Datum q = HnswGetValue(base, element);
HnswElement skipElement = existing ? element : NULL;
bool inMemory = index == NULL;
q.value = HnswGetValue(base, element);
/* Precompute hash */
if (inMemory)
if (index == NULL)
PrecomputeHash(base, element);
/* No neighbors if no entry point */
@@ -1297,13 +1268,13 @@ HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint
return;
/* Get entry point and level */
ep = list_make1(HnswEntryCandidate(base, entryPoint, &q, index, support, true));
ep = list_make1(HnswEntryCandidate(base, entryPoint, q, index, procinfo, collation, true));
entryLevel = entryPoint->level;
/* 1st phase: greedy search to insert level */
for (int lc = entryLevel; lc >= level + 1; lc--)
{
w = HnswSearchLayer(base, &q, ep, 1, lc, index, support, m, true, skipElement, NULL, NULL, true, NULL);
w = HnswSearchLayer(base, q, ep, 1, lc, index, procinfo, collation, m, true, skipElement);
ep = w;
}
@@ -1322,7 +1293,7 @@ HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint
List *lw = NIL;
ListCell *lc2;
w = HnswSearchLayer(base, &q, ep, efConstruction, lc, index, support, m, true, skipElement, NULL, NULL, true, NULL);
w = HnswSearchLayer(base, q, ep, efConstruction, lc, index, procinfo, collation, m, true, skipElement);
/* Convert search candidates to candidates */
foreach(lc2, w)
@@ -1338,7 +1309,7 @@ HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint
/* Elements being deleted or skipped can help with search */
/* but should be removed before selecting neighbors */
if (!inMemory)
if (index != NULL)
lw = RemoveElements(base, lw, skipElement);
/*
@@ -1346,7 +1317,7 @@ HnswFindElementNeighbors(char *base, HnswElement element, HnswElement entryPoint
* sortCandidates to true for in-memory builds to enable closer
* caching, but there does not seem to be a difference in performance.
*/
neighbors = SelectNeighbors(base, lw, lm, support, &HnswGetNeighbors(base, element, lc)->closerSet, NULL, NULL, false);
neighbors = SelectNeighbors(base, lw, lm, lc, procinfo, collation, element, NULL, NULL, false);
AddConnections(base, element, neighbors, lc);
@@ -1404,19 +1375,6 @@ hnsw_halfvec_support(PG_FUNCTION_ARGS)
PG_RETURN_POINTER(&typeInfo);
};
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(hnsw_intvec_support);
Datum
hnsw_intvec_support(PG_FUNCTION_ARGS)
{
static const HnswTypeInfo typeInfo = {
.maxDimensions = HNSW_MAX_DIM * 4,
.normalize = NULL,
.checkValue = NULL
};
PG_RETURN_POINTER(&typeInfo);
};
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(hnsw_bit_support);
Datum
hnsw_bit_support(PG_FUNCTION_ARGS)

View File

@@ -184,12 +184,13 @@ static void
RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswElement entryPoint)
{
Relation index = vacuumstate->index;
HnswSupport *support = &vacuumstate->support;
Buffer buf;
Page page;
GenericXLogState *state;
int m = vacuumstate->m;
int efConstruction = vacuumstate->efConstruction;
FmgrInfo *procinfo = vacuumstate->procinfo;
Oid collation = vacuumstate->collation;
BufferAccessStrategy bas = vacuumstate->bas;
HnswNeighborTuple ntup = vacuumstate->ntup;
Size ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m);
@@ -204,7 +205,7 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
element->heaptidsLength = 0;
/* Find neighbors for element, skipping itself */
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, true);
HnswFindElementNeighbors(base, element, entryPoint, index, procinfo, collation, m, efConstruction, true);
/* Zero memory for each element */
MemSet(ntup, 0, HNSW_TUPLE_ALLOC_SIZE);
@@ -228,7 +229,7 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
UnlockReleaseBuffer(buf);
/* Update neighbors */
HnswUpdateNeighborsOnDisk(index, support, element, m, true, false);
HnswUpdateNeighborsOnDisk(index, procinfo, collation, element, m, true, false);
}
/*
@@ -238,7 +239,6 @@ static void
RepairGraphEntryPoint(HnswVacuumState * vacuumstate)
{
Relation index = vacuumstate->index;
HnswSupport *support = &vacuumstate->support;
HnswElement highestPoint = &vacuumstate->highestPoint;
HnswElement entryPoint;
MemoryContext oldCtx = MemoryContextSwitchTo(vacuumstate->tmpCtx);
@@ -256,7 +256,7 @@ RepairGraphEntryPoint(HnswVacuumState * vacuumstate)
LockPage(index, HNSW_UPDATE_LOCK, ShareLock);
/* Load element */
HnswLoadElement(highestPoint, NULL, NULL, index, support, true, NULL);
HnswLoadElement(highestPoint, NULL, NULL, index, vacuumstate->procinfo, vacuumstate->collation, true, NULL);
/* Repair if needed */
if (NeedsUpdated(vacuumstate, highestPoint))
@@ -294,7 +294,7 @@ RepairGraphEntryPoint(HnswVacuumState * vacuumstate)
* is outdated, this can remove connections at higher levels in
* the graph until they are repaired, but this should be fine.
*/
HnswLoadElement(entryPoint, NULL, NULL, index, support, true, NULL);
HnswLoadElement(entryPoint, NULL, NULL, index, vacuumstate->procinfo, vacuumstate->collation, true, NULL);
if (NeedsUpdated(vacuumstate, entryPoint))
{
@@ -527,14 +527,6 @@ MarkDeleted(HnswVacuumState * vacuumstate)
for (int i = 0; i < ntup->count; i++)
ItemPointerSetInvalid(&ntup->indextids[i]);
/* Increment version */
/* This is used to avoid incorrect reads for iterative scans */
/* Reserve some bits for future use */
etup->version++;
if (etup->version > 15)
etup->version = 1;
ntup->version = etup->version;
/*
* We modified the tuples in place, no need to call
* PageIndexTupleOverwrite
@@ -581,13 +573,13 @@ InitVacuumState(HnswVacuumState * vacuumstate, IndexVacuumInfo *info, IndexBulkD
vacuumstate->callback_state = callback_state;
vacuumstate->efConstruction = HnswGetEfConstruction(index);
vacuumstate->bas = GetAccessStrategy(BAS_BULKREAD);
vacuumstate->procinfo = index_getprocinfo(index, 1, HNSW_DISTANCE_PROC);
vacuumstate->collation = index->rd_indcollation[0];
vacuumstate->ntup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
vacuumstate->tmpCtx = AllocSetContextCreate(CurrentMemoryContext,
"Hnsw vacuum temporary context",
ALLOCSET_DEFAULT_SIZES);
HnswInitSupport(&vacuumstate->support, index);
/* Get m from metapage */
HnswGetMetaPageInfo(index, &vacuumstate->m, NULL);

View File

@@ -1,754 +0,0 @@
#include "postgres.h"
#include <limits.h>
#include <math.h>
#include "catalog/pg_type.h"
#include "fmgr.h"
#include "intvec.h"
#include "lib/stringinfo.h"
#include "libpq/pqformat.h"
#include "utils/array.h"
#include "utils/builtins.h"
#include "utils/lsyscache.h"
/*
* Ensure same dimensions
*/
static inline void
CheckDims(IntVector * a, IntVector * b)
{
if (a->dim != b->dim)
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("different intvec dimensions %d and %d", a->dim, b->dim)));
}
/*
* Ensure expected dimensions
*/
static inline void
CheckExpectedDim(int32 typmod, int dim)
{
if (typmod != -1 && typmod != dim)
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("expected %d dimensions, not %d", typmod, dim)));
}
/*
* Ensure valid dimensions
*/
static inline void
CheckDim(int dim)
{
if (dim < 1)
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("intvec must have at least 1 dimension")));
if (dim > INTVEC_MAX_DIM)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("intvec cannot have more than %d dimensions", INTVEC_MAX_DIM)));
}
/*
* Ensure element in range
*/
static inline void
CheckElement(long value)
{
if (value < SCHAR_MIN || value > SCHAR_MAX)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("value \"%ld\" is out of range for type intvec", value)));
}
/*
* Allocate and initialize a new int vector
*/
IntVector *
InitIntVector(int dim)
{
IntVector *result;
int size;
size = INTVEC_SIZE(dim);
result = (IntVector *) palloc0(size);
SET_VARSIZE(result, size);
result->dim = dim;
return result;
}
/*
* Check for whitespace, since array_isspace() is static
*/
static inline bool
intvec_isspace(char ch)
{
if (ch == ' ' ||
ch == '\t' ||
ch == '\n' ||
ch == '\r' ||
ch == '\v' ||
ch == '\f')
return true;
return false;
}
/*
* Convert textual representation to internal representation
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_in);
Datum
intvec_in(PG_FUNCTION_ARGS)
{
char *lit = PG_GETARG_CSTRING(0);
int32 typmod = PG_GETARG_INT32(2);
int8 x[INTVEC_MAX_DIM];
int dim = 0;
char *pt = lit;
IntVector *result;
while (intvec_isspace(*pt))
pt++;
if (*pt != '[')
ereport(ERROR,
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
errmsg("invalid input syntax for type intvec: \"%s\"", lit),
errdetail("Vector contents must start with \"[\".")));
pt++;
while (intvec_isspace(*pt))
pt++;
if (*pt == ']')
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("intvec must have at least 1 dimension")));
for (;;)
{
long val;
char *stringEnd;
if (dim == INTVEC_MAX_DIM)
ereport(ERROR,
(errcode(ERRCODE_PROGRAM_LIMIT_EXCEEDED),
errmsg("intvec cannot have more than %d dimensions", VECTOR_MAX_DIM)));
while (intvec_isspace(*pt))
pt++;
/* Check for empty string like float4in */
if (*pt == '\0')
ereport(ERROR,
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
errmsg("invalid input syntax for type intvec: \"%s\"", lit)));
errno = 0;
/* Use similar logic as int2vectorin */
val = strtol(pt, &stringEnd, 10);
if (stringEnd == pt)
ereport(ERROR,
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
errmsg("invalid input syntax for type intvec: \"%s\"", lit)));
/* Check for range error like float4in */
if (errno == ERANGE || val < SCHAR_MIN || val > SCHAR_MAX)
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("\"%s\" is out of range for type intvec", pnstrdup(pt, stringEnd - pt))));
CheckElement(val);
x[dim++] = val;
pt = stringEnd;
while (intvec_isspace(*pt))
pt++;
if (*pt == ',')
pt++;
else if (*pt == ']')
{
pt++;
break;
}
else
ereport(ERROR,
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
errmsg("invalid input syntax for type intvec: \"%s\"", lit)));
}
/* Only whitespace is allowed after the closing brace */
while (intvec_isspace(*pt))
pt++;
if (*pt != '\0')
ereport(ERROR,
(errcode(ERRCODE_INVALID_TEXT_REPRESENTATION),
errmsg("invalid input syntax for type intvec: \"%s\"", lit),
errdetail("Junk after closing right brace.")));
CheckDim(dim);
CheckExpectedDim(typmod, dim);
result = InitIntVector(dim);
for (int i = 0; i < dim; i++)
result->x[i] = x[i];
PG_RETURN_POINTER(result);
}
/*
* Convert internal representation to textual representation
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_out);
Datum
intvec_out(PG_FUNCTION_ARGS)
{
IntVector *vector = PG_GETARG_INTVEC_P(0);
int dim = vector->dim;
char *buf;
char *ptr;
/*
* Need:
*
* dim * 4 bytes for elements (-128 to 127)
*
* dim - 1 bytes for separator
*
* 3 bytes for [, ], and \0
*/
buf = (char *) palloc(5 * dim + 2);
ptr = buf;
*ptr = '[';
ptr++;
for (int i = 0; i < dim; i++)
{
if (i > 0)
{
*ptr = ',';
ptr++;
}
#if PG_VERSION_NUM >= 140000
ptr += pg_ltoa(vector->x[i], ptr);
#else
pg_ltoa(vector->x[i], ptr);
while (*ptr != '\0')
ptr++;
#endif
}
*ptr = ']';
ptr++;
*ptr = '\0';
PG_FREE_IF_COPY(vector, 0);
PG_RETURN_CSTRING(buf);
}
/*
* Convert type modifier
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_typmod_in);
Datum
intvec_typmod_in(PG_FUNCTION_ARGS)
{
ArrayType *ta = PG_GETARG_ARRAYTYPE_P(0);
int32 *tl;
int n;
tl = ArrayGetIntegerTypmods(ta, &n);
if (n != 1)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid type modifier")));
if (*tl < 1)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("dimensions for type intvec must be at least 1")));
if (*tl > INTVEC_MAX_DIM)
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("dimensions for type intvec cannot exceed %d", INTVEC_MAX_DIM)));
PG_RETURN_INT32(*tl);
}
/*
* Convert external binary representation to internal representation
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_recv);
Datum
intvec_recv(PG_FUNCTION_ARGS)
{
StringInfo buf = (StringInfo) PG_GETARG_POINTER(0);
int32 typmod = PG_GETARG_INT32(2);
IntVector *result;
int16 dim;
int16 unused;
dim = pq_getmsgint(buf, sizeof(int16));
unused = pq_getmsgint(buf, sizeof(int16));
CheckDim(dim);
CheckExpectedDim(typmod, dim);
if (unused != 0)
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("expected unused to be 0, not %d", unused)));
result = InitIntVector(dim);
for (int i = 0; i < dim; i++)
result->x[i] = pq_getmsgint(buf, sizeof(int8));
PG_RETURN_POINTER(result);
}
/*
* Convert internal representation to the external binary representation
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_send);
Datum
intvec_send(PG_FUNCTION_ARGS)
{
IntVector *vec = PG_GETARG_INTVEC_P(0);
StringInfoData buf;
pq_begintypsend(&buf);
pq_sendint(&buf, vec->dim, sizeof(int16));
pq_sendint(&buf, vec->unused, sizeof(int16));
for (int i = 0; i < vec->dim; i++)
pq_sendint8(&buf, vec->x[i]);
PG_RETURN_BYTEA_P(pq_endtypsend(&buf));
}
/*
* Convert int vector to int vector
* This is needed to check the type modifier
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec);
Datum
intvec(PG_FUNCTION_ARGS)
{
IntVector *vec = PG_GETARG_INTVEC_P(0);
int32 typmod = PG_GETARG_INT32(1);
CheckExpectedDim(typmod, vec->dim);
PG_RETURN_POINTER(vec);
}
/*
* Convert array to intvec vector
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(array_to_intvec);
Datum
array_to_intvec(PG_FUNCTION_ARGS)
{
ArrayType *array = PG_GETARG_ARRAYTYPE_P(0);
int32 typmod = PG_GETARG_INT32(1);
IntVector *result;
int16 typlen;
bool typbyval;
char typalign;
Datum *elemsp;
int nelemsp;
if (ARR_NDIM(array) > 1)
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("array must be 1-D")));
if (ARR_HASNULL(array) && array_contains_nulls(array))
ereport(ERROR,
(errcode(ERRCODE_NULL_VALUE_NOT_ALLOWED),
errmsg("array must not contain nulls")));
get_typlenbyvalalign(ARR_ELEMTYPE(array), &typlen, &typbyval, &typalign);
deconstruct_array(array, ARR_ELEMTYPE(array), typlen, typbyval, typalign, &elemsp, NULL, &nelemsp);
CheckDim(nelemsp);
CheckExpectedDim(typmod, nelemsp);
result = InitIntVector(nelemsp);
if (ARR_ELEMTYPE(array) == INT4OID)
{
for (int i = 0; i < nelemsp; i++)
{
long l = DatumGetInt32(elemsp[i]);
CheckElement(l);
result->x[i] = l;
}
}
else
{
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("unsupported array type")));
}
/*
* Free allocation from deconstruct_array. Do not free individual elements
* when pass-by-reference since they point to original array.
*/
pfree(elemsp);
PG_RETURN_POINTER(result);
}
/*
* Convert int vector to integer[]
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_to_integer);
Datum
intvec_to_integer(PG_FUNCTION_ARGS)
{
IntVector *vec = PG_GETARG_INTVEC_P(0);
Datum *datums;
ArrayType *result;
datums = (Datum *) palloc(sizeof(Datum) * vec->dim);
for (int i = 0; i < vec->dim; i++)
datums[i] = Int32GetDatum((int) vec->x[i]);
result = construct_array(datums, vec->dim, INT4OID, sizeof(int32), true, TYPALIGN_INT);
pfree(datums);
PG_RETURN_POINTER(result);
}
static int
IntvecL2SquaredDistance(int dim, int8 *ax, int8 *bx)
{
int distance = 0;
/* Auto-vectorized */
for (int i = 0; i < dim; i++)
{
int diff = (int) ax[i] - (int) bx[i];
distance += diff * diff;
}
return distance;
}
/*
* Get the L2 distance between int vectors
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l2_distance);
Datum
intvec_l2_distance(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
CheckDims(a, b);
PG_RETURN_FLOAT8(sqrt((double) IntvecL2SquaredDistance(a->dim, a->x, b->x)));
}
/*
* Get the L2 squared distance between int vectors
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l2_squared_distance);
Datum
intvec_l2_squared_distance(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
CheckDims(a, b);
PG_RETURN_FLOAT8((double) IntvecL2SquaredDistance(a->dim, a->x, b->x));
}
static int
IntvecInnerProduct(int dim, int8 *ax, int8 *bx)
{
int distance = 0;
/* Auto-vectorized */
for (int i = 0; i < dim; i++)
distance += (int) ax[i] * (int) bx[i];
return distance;
}
/*
* Get the inner product of two int vectors
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_inner_product);
Datum
intvec_inner_product(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
CheckDims(a, b);
PG_RETURN_FLOAT8((double) IntvecInnerProduct(a->dim, a->x, b->x));
}
/*
* Get the negative inner product of two int vectors
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_negative_inner_product);
Datum
intvec_negative_inner_product(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
CheckDims(a, b);
PG_RETURN_FLOAT8((double) -IntvecInnerProduct(a->dim, a->x, b->x));
}
static double
IntvecCosineSimilarity(int dim, int8 *ax, int8 *bx)
{
int similarity = 0;
int norma = 0;
int normb = 0;
/* Auto-vectorized */
for (int i = 0; i < dim; i++)
{
int axi = ax[i];
int bxi = bx[i];
similarity += axi * bxi;
norma += axi * axi;
normb += bxi * bxi;
}
/* Use sqrt(a * b) over sqrt(a) * sqrt(b) */
return (double) similarity / sqrt((double) norma * (double) normb);
}
/*
* Get the cosine distance between two int vectors
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_cosine_distance);
Datum
intvec_cosine_distance(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
double similarity;
CheckDims(a, b);
similarity = IntvecCosineSimilarity(a->dim, a->x, b->x);
#ifdef _MSC_VER
/* /fp:fast may not propagate NaN */
if (isnan(similarity))
PG_RETURN_FLOAT8(NAN);
#endif
/* Keep in range */
if (similarity > 1)
similarity = 1;
else if (similarity < -1)
similarity = -1;
PG_RETURN_FLOAT8(1 - similarity);
}
static int
IntvecL1Distance(int dim, int8 *ax, int8 *bx)
{
int distance = 0;
/* Auto-vectorized */
for (int i = 0; i < dim; i++)
distance += abs((int) ax[i] - (int) bx[i]);
return distance;
}
/*
* Get the L1 distance between two int vectors
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l1_distance);
Datum
intvec_l1_distance(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
CheckDims(a, b);
PG_RETURN_FLOAT8((double) IntvecL1Distance(a->dim, a->x, b->x));
}
/*
* Get the dimensions of an int vector
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_vector_dims);
Datum
intvec_vector_dims(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
PG_RETURN_INT32(a->dim);
}
/*
* Get the L2 norm of an int vector
*/
PGDLLEXPORT PG_FUNCTION_INFO_V1(intvec_l2_norm);
Datum
intvec_l2_norm(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
int8 *ax = a->x;
int norm = 0;
/* Auto-vectorized */
for (int i = 0; i < a->dim; i++)
norm += (int) ax[i] * (int) ax[i];
PG_RETURN_FLOAT8(sqrt((double) norm));
}
/*
* Internal helper to compare int vectors
*/
static int
intvec_cmp_internal(IntVector * a, IntVector * b)
{
int dim = Min(a->dim, b->dim);
/* Check values before dimensions to be consistent with Postgres arrays */
for (int i = 0; i < dim; i++)
{
if (a->x[i] < b->x[i])
return -1;
if (a->x[i] > b->x[i])
return 1;
}
if (a->dim < b->dim)
return -1;
if (a->dim > b->dim)
return 1;
return 0;
}
/*
* Less than
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_lt);
Datum
intvec_lt(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_BOOL(intvec_cmp_internal(a, b) < 0);
}
/*
* Less than or equal
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_le);
Datum
intvec_le(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_BOOL(intvec_cmp_internal(a, b) <= 0);
}
/*
* Equal
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_eq);
Datum
intvec_eq(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_BOOL(intvec_cmp_internal(a, b) == 0);
}
/*
* Not equal
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_ne);
Datum
intvec_ne(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_BOOL(intvec_cmp_internal(a, b) != 0);
}
/*
* Greater than or equal
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_ge);
Datum
intvec_ge(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_BOOL(intvec_cmp_internal(a, b) >= 0);
}
/*
* Greater than
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_gt);
Datum
intvec_gt(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_BOOL(intvec_cmp_internal(a, b) > 0);
}
/*
* Compare int vectors
*/
FUNCTION_PREFIX PG_FUNCTION_INFO_V1(intvec_cmp);
Datum
intvec_cmp(PG_FUNCTION_ARGS)
{
IntVector *a = PG_GETARG_INTVEC_P(0);
IntVector *b = PG_GETARG_INTVEC_P(1);
PG_RETURN_INT32(intvec_cmp_internal(a, b));
}

View File

@@ -1,23 +0,0 @@
#ifndef INTVEC_H
#define INTVEC_H
#include "vector.h"
#define INTVEC_MAX_DIM VECTOR_MAX_DIM
#define INTVEC_SIZE(_dim) (offsetof(IntVector, x) + sizeof(int8)*(_dim))
#define DatumGetIntVector(x) ((IntVector *) PG_DETOAST_DATUM(x))
#define PG_GETARG_INTVEC_P(x) DatumGetIntVector(PG_GETARG_DATUM(x))
#define PG_RETURN_INTVEC_P(x) PG_RETURN_POINTER(x)
typedef struct IntVector
{
int32 vl_len_; /* varlena header (do not touch directly!) */
int16 dim; /* number of dimensions */
int16 unused;
int8 x[FLEXIBLE_ARRAY_MEMBER];
} IntVector;
IntVector *InitIntVector(int dim);
#endif

View File

@@ -228,11 +228,11 @@ BuildCallback(Relation index, ItemPointer tid, Datum *values,
static inline void
GetNextTuple(Tuplesortstate *sortstate, TupleDesc tupdesc, TupleTableSlot *slot, IndexTuple *itup, int *list)
{
Datum value;
bool isnull;
if (tuplesort_gettupleslot(sortstate, true, false, slot, NULL))
{
Datum value;
bool isnull;
*list = DatumGetInt32(slot_getattr(slot, 1, &isnull));
value = slot_getattr(slot, 3, &isnull);
@@ -254,8 +254,8 @@ InsertTuples(Relation index, IvfflatBuildState * buildstate, ForkNumber forkNum)
IndexTuple itup = NULL; /* silence compiler warning */
int64 inserted = 0;
TupleTableSlot *slot = MakeSingleTupleTableSlot(buildstate->sortdesc, &TTSOpsMinimalTuple);
TupleDesc tupdesc = buildstate->tupdesc;
TupleTableSlot *slot = MakeSingleTupleTableSlot(buildstate->tupdesc, &TTSOpsMinimalTuple);
TupleDesc tupdesc = RelationGetDescr(index);
pgstat_progress_update_param(PROGRESS_CREATEIDX_SUBPHASE, PROGRESS_IVFFLAT_PHASE_LOAD);
@@ -319,7 +319,6 @@ InitBuildState(IvfflatBuildState * buildstate, Relation heap, Relation index, In
buildstate->index = index;
buildstate->indexInfo = indexInfo;
buildstate->typeInfo = IvfflatGetTypeInfo(index);
buildstate->tupdesc = RelationGetDescr(index);
buildstate->lists = IvfflatGetLists(index);
buildstate->dimensions = TupleDescAttr(index->rd_att, 0)->atttypmod;
@@ -357,12 +356,12 @@ InitBuildState(IvfflatBuildState * buildstate, Relation heap, Relation index, In
errmsg("dimensions must be greater than one for this opclass")));
/* Create tuple description for sorting */
buildstate->sortdesc = CreateTemplateTupleDesc(3);
TupleDescInitEntry(buildstate->sortdesc, (AttrNumber) 1, "list", INT4OID, -1, 0);
TupleDescInitEntry(buildstate->sortdesc, (AttrNumber) 2, "tid", TIDOID, -1, 0);
TupleDescInitEntry(buildstate->sortdesc, (AttrNumber) 3, "vector", buildstate->tupdesc->attrs[0].atttypid, -1, 0);
buildstate->tupdesc = CreateTemplateTupleDesc(3);
TupleDescInitEntry(buildstate->tupdesc, (AttrNumber) 1, "list", INT4OID, -1, 0);
TupleDescInitEntry(buildstate->tupdesc, (AttrNumber) 2, "tid", TIDOID, -1, 0);
TupleDescInitEntry(buildstate->tupdesc, (AttrNumber) 3, "vector", RelationGetDescr(index)->attrs[0].atttypid, -1, 0);
buildstate->slot = MakeSingleTupleTableSlot(buildstate->sortdesc, &TTSOpsVirtual);
buildstate->slot = MakeSingleTupleTableSlot(buildstate->tupdesc, &TTSOpsVirtual);
buildstate->centers = VectorArrayInit(buildstate->lists, buildstate->dimensions, buildstate->typeInfo->itemSize(buildstate->dimensions));
buildstate->listInfo = palloc(sizeof(ListInfo) * buildstate->lists);
@@ -634,7 +633,7 @@ IvfflatParallelScanAndSort(IvfflatSpool * ivfspool, IvfflatShared * ivfshared, S
InitBuildState(&buildstate, ivfspool->heap, ivfspool->index, indexInfo);
memcpy(buildstate.centers->items, ivfcenters, buildstate.centers->itemsize * buildstate.centers->maxlen);
buildstate.centers->length = buildstate.centers->maxlen;
ivfspool->sortstate = InitBuildSortState(buildstate.sortdesc, sortmem, coordinate);
ivfspool->sortstate = InitBuildSortState(buildstate.tupdesc, sortmem, coordinate);
buildstate.sortstate = ivfspool->sortstate;
scan = table_beginscan_parallel(ivfspool->heap,
ParallelTableScanFromIvfflatShared(ivfshared));
@@ -951,7 +950,7 @@ AssignTuples(IvfflatBuildState * buildstate)
}
/* Begin serial/leader tuplesort */
buildstate->sortstate = InitBuildSortState(buildstate->sortdesc, maintenance_work_mem, coordinate);
buildstate->sortstate = InitBuildSortState(buildstate->tupdesc, maintenance_work_mem, coordinate);
/* Add tuples to sort */
if (buildstate->heap != NULL)

View File

@@ -17,16 +17,10 @@
#endif
int ivfflat_probes;
int ivfflat_iterative_search;
int ivfflat_max_probes;
bool ivfflat_streaming;
static relopt_kind ivfflat_relopt_kind;
static const struct config_enum_entry ivfflat_iterative_search_options[] = {
{"off", IVFFLAT_ITERATIVE_SEARCH_OFF, false},
{"relaxed_order", IVFFLAT_ITERATIVE_SEARCH_RELAXED, false},
{NULL, 0, false}
};
/*
* Initialize index options and variables
*/
@@ -41,14 +35,13 @@ IvfflatInit(void)
"Valid range is 1..lists.", &ivfflat_probes,
IVFFLAT_DEFAULT_PROBES, IVFFLAT_MIN_LISTS, IVFFLAT_MAX_LISTS, PGC_USERSET, 0, NULL, NULL, NULL);
DefineCustomEnumVariable("ivfflat.iterative_search", "Sets the iterative search mode",
NULL, &ivfflat_iterative_search,
IVFFLAT_ITERATIVE_SEARCH_OFF, ivfflat_iterative_search_options, PGC_USERSET, 0, NULL, NULL, NULL);
DefineCustomIntVariable("ivfflat.max_probes", "Sets the max number of probes for iterative scans",
NULL, &ivfflat_max_probes,
-1, IVFFLAT_MIN_LISTS, IVFFLAT_MAX_LISTS, PGC_USERSET, 0, NULL, NULL, NULL);
/* If this is less than probes, probes is used */
DefineCustomIntVariable("ivfflat.max_probes", "Sets the max number of probes for iterative search",
"-1 means no limit", &ivfflat_max_probes,
-1, -1, IVFFLAT_MAX_LISTS, PGC_USERSET, 0, NULL, NULL, NULL);
DefineCustomBoolVariable("ivfflat.streaming", "Use streaming mode",
NULL, &ivfflat_streaming,
IVFFLAT_DEFAULT_STREAMING, PGC_USERSET, 0, NULL, NULL, NULL);
MarkGUCPrefixReserved("ivfflat");
}
@@ -74,6 +67,35 @@ ivfflatbuildphasename(int64 phasenum)
}
}
/*
* Estimate the number of probes for iterative scans
*/
static int
EstimateProbes(PlannerInfo *root, IndexPath *path, int lists)
{
double selectivity = 1;
ListCell *lc;
double tuplesPerList;
/* Cannot estimate without limit */
/* limit_tuples includes offset */
if (root->limit_tuples < 0)
return 0;
/* Get the selectivity of non-index conditions */
foreach(lc, path->indexinfo->indrestrictinfo)
{
RestrictInfo *rinfo = lfirst(lc);
/* Skip DEFAULT_INEQ_SEL since it may be a distance filter */
if (rinfo->norm_selec >= 0 && rinfo->norm_selec <= 1 && rinfo->norm_selec != (Selectivity) DEFAULT_INEQ_SEL)
selectivity *= rinfo->norm_selec;
}
tuplesPerList = path->indexinfo->tuples / (double) lists;
return root->limit_tuples / (tuplesPerList * selectivity);
}
/*
* Estimate the cost of an index scan
*/
@@ -85,9 +107,8 @@ ivfflatcostestimate(PlannerInfo *root, IndexPath *path, double loop_count,
{
GenericCosts costs;
int lists;
int probes;
double ratio;
double sequentialRatio = 0.5;
double startupPages;
double spc_seq_page_cost;
Relation index;
@@ -104,37 +125,59 @@ ivfflatcostestimate(PlannerInfo *root, IndexPath *path, double loop_count,
MemSet(&costs, 0, sizeof(costs));
genericcostestimate(root, path, loop_count, &costs);
index = index_open(path->indexinfo->indexoid, NoLock);
IvfflatGetMetaPageInfo(index, &lists, NULL);
index_close(index, NoLock);
probes = ivfflat_probes;
if (ivfflat_streaming)
{
probes = Max(probes, EstimateProbes(root, path, lists));
if (ivfflat_max_probes != -1)
probes = Min(probes, ivfflat_max_probes);
}
/* Get the ratio of lists that we need to visit */
ratio = ((double) ivfflat_probes) / lists;
ratio = ((double) probes) / lists;
if (ratio > 1.0)
ratio = 1.0;
/*
* This gives us the subset of tuples to visit. This value is passed into
* the generic cost estimator to determine the number of pages to visit
* during the index scan.
*/
costs.numIndexTuples = path->indexinfo->tuples * ratio;
genericcostestimate(root, path, loop_count, &costs);
get_tablespace_page_costs(path->indexinfo->reltablespace, NULL, &spc_seq_page_cost);
/* Change some page cost from random to sequential */
costs.indexTotalCost -= sequentialRatio * costs.numIndexPages * (costs.spc_random_page_cost - spc_seq_page_cost);
/* Startup cost is cost before returning the first row */
costs.indexStartupCost = costs.indexTotalCost * ratio;
/* Adjust cost if needed since TOAST not included in seq scan cost */
startupPages = costs.numIndexPages * ratio;
if (startupPages > path->indexinfo->rel->pages && ratio < 0.5)
if (costs.numIndexPages > path->indexinfo->rel->pages && ratio < 0.5)
{
/* Change rest of page cost from random to sequential */
costs.indexStartupCost -= (1 - sequentialRatio) * startupPages * (costs.spc_random_page_cost - spc_seq_page_cost);
/* Change all page cost from random to sequential */
costs.indexTotalCost -= costs.numIndexPages * (costs.spc_random_page_cost - spc_seq_page_cost);
/* Remove cost of extra pages */
costs.indexStartupCost -= (startupPages - path->indexinfo->rel->pages) * spc_seq_page_cost;
costs.indexTotalCost -= (costs.numIndexPages - path->indexinfo->rel->pages) * spc_seq_page_cost;
}
else
{
/* Change some page cost from random to sequential */
costs.indexTotalCost -= 0.5 * costs.numIndexPages * (costs.spc_random_page_cost - spc_seq_page_cost);
}
*indexStartupCost = costs.indexStartupCost;
/*
* If the list selectivity is lower than what is returned from the generic
* cost estimator, use that.
*/
if (ratio < costs.indexSelectivity)
costs.indexSelectivity = ratio;
/* Use total cost since most work happens before first tuple is returned */
*indexStartupCost = costs.indexTotalCost;
*indexTotalCost = costs.indexTotalCost;
*indexSelectivity = costs.indexSelectivity;
*indexCorrelation = costs.indexCorrelation;

View File

@@ -43,6 +43,7 @@
#define IVFFLAT_MIN_LISTS 1
#define IVFFLAT_MAX_LISTS 32768
#define IVFFLAT_DEFAULT_PROBES 1
#define IVFFLAT_DEFAULT_STREAMING false
/* Build phases */
/* PROGRESS_CREATEIDX_SUBPHASE_INITIALIZE is 1 */
@@ -80,14 +81,8 @@
/* Variables */
extern int ivfflat_probes;
extern int ivfflat_iterative_search;
extern int ivfflat_max_probes;
typedef enum IvfflatIterativeSearchMode
{
IVFFLAT_ITERATIVE_SEARCH_OFF,
IVFFLAT_ITERATIVE_SEARCH_RELAXED
} IvfflatIterativeSearchMode;
extern bool ivfflat_streaming;
typedef struct VectorArrayData
{
@@ -173,7 +168,6 @@ typedef struct IvfflatBuildState
Relation index;
IndexInfo *indexInfo;
const IvfflatTypeInfo *typeInfo;
TupleDesc tupdesc;
/* Settings */
int dimensions;
@@ -207,7 +201,7 @@ typedef struct IvfflatBuildState
/* Sorting */
Tuplesortstate *sortstate;
TupleDesc sortdesc;
TupleDesc tupdesc;
TupleTableSlot *slot;
/* Memory */
@@ -260,7 +254,6 @@ typedef struct IvfflatScanOpaqueData
int dimensions;
bool first;
Datum value;
MemoryContext tmpCtx;
/* Sorting */
Tuplesortstate *sortstate;
@@ -277,9 +270,9 @@ typedef struct IvfflatScanOpaqueData
/* Lists */
pairingheap *listQueue;
BlockNumber *listPages;
int listIndex;
IvfflatScanList *lists;
BlockNumber *startPages;
int currentIndex;
IvfflatScanList lists[FLEXIBLE_ARRAY_MEMBER]; /* must come last */
} IvfflatScanOpaqueData;
typedef IvfflatScanOpaqueData * IvfflatScanOpaque;

View File

@@ -98,7 +98,7 @@ InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heap_tid, R
IvfflatGetMetaPageInfo(index, NULL, NULL);
/* Find the insert page - sets the page and list info */
FindInsertPage(index, &value, &insertPage, &listInfo);
FindInsertPage(index, values, &insertPage, &listInfo);
Assert(BlockNumberIsValid(insertPage));
originalInsertPage = insertPage;

View File

@@ -10,7 +10,10 @@
#include "miscadmin.h"
#include "pgstat.h"
#include "storage/bufmgr.h"
#ifdef IVFFLAT_MEMORY
#include "utils/memutils.h"
#endif
#define GetScanList(ptr) pairingheap_container(IvfflatScanList, ph_node, ptr)
#define GetScanListConst(ptr) pairingheap_const_container(IvfflatScanList, ph_node, ptr)
@@ -75,7 +78,7 @@ GetScanLists(IndexScanDesc scan, Datum value)
pairingheap_add(so->listQueue, &scanlist->ph_node);
/* Calculate max distance */
if (listCount == so->maxProbes)
if (listCount == so->probes)
maxDistance = GetScanList(pairingheap_first(so->listQueue))->distance;
}
else if (distance < maxDistance)
@@ -101,7 +104,7 @@ GetScanLists(IndexScanDesc scan, Datum value)
}
for (int i = listCount - 1; i >= 0; i--)
so->listPages[i] = GetScanList(pairingheap_remove_first(so->listQueue))->startPage;
so->startPages[i] = GetScanList(pairingheap_remove_first(so->listQueue))->startPage;
Assert(pairingheap_is_empty(so->listQueue));
}
@@ -121,9 +124,9 @@ GetScanItems(IndexScanDesc scan, Datum value)
tuplesort_reset(so->sortstate);
/* Search closest probes lists */
while (so->listIndex < so->maxProbes && (++batchProbes) <= so->probes)
while (so->currentIndex < so->maxProbes && (++batchProbes) <= so->probes)
{
BlockNumber searchPage = so->listPages[so->listIndex++];
BlockNumber searchPage = so->startPages[so->currentIndex++];
/* Search all entry pages for list */
while (BlockNumberIsValid(searchPage))
@@ -171,7 +174,7 @@ GetScanItems(IndexScanDesc scan, Datum value)
}
}
if (tuples < 100 && ivfflat_iterative_search == IVFFLAT_ITERATIVE_SEARCH_OFF)
if (tuples < 100 && !ivfflat_streaming)
ereport(DEBUG1,
(errmsg("index scan found few tuples"),
errdetail("Index may have been created with little data."),
@@ -218,13 +221,7 @@ GetScanValue(IndexScanDesc scan)
/* Normalize if needed */
if (so->normprocinfo != NULL)
{
MemoryContext oldCtx = MemoryContextSwitchTo(so->tmpCtx);
value = IvfflatNormValue(so->typeInfo, so->collation, value);
MemoryContextSwitchTo(oldCtx);
}
}
return value;
@@ -256,35 +253,26 @@ ivfflatbeginscan(Relation index, int nkeys, int norderbys)
int dimensions;
int probes = ivfflat_probes;
int maxProbes;
MemoryContext oldCtx;
scan = RelationGetIndexScan(index, nkeys, norderbys);
/* Get lists and dimensions from metapage */
IvfflatGetMetaPageInfo(index, &lists, &dimensions);
if (ivfflat_iterative_search != IVFFLAT_ITERATIVE_SEARCH_OFF)
{
maxProbes = ivfflat_max_probes;
if (probes > lists)
probes = lists;
if (maxProbes < 0)
if (ivfflat_streaming)
{
if (ivfflat_max_probes == -1)
maxProbes = lists;
else if (maxProbes < probes)
{
/* TODO Show notice */
maxProbes = probes;
}
else
maxProbes = Min(ivfflat_max_probes, lists);
}
else
maxProbes = probes;
if (probes > lists)
probes = lists;
if (maxProbes > lists)
maxProbes = lists;
so = (IvfflatScanOpaque) palloc(sizeof(IvfflatScanOpaqueData));
so = (IvfflatScanOpaque) palloc(offsetof(IvfflatScanOpaqueData, lists) + maxProbes * sizeof(IvfflatScanList));
so->typeInfo = IvfflatGetTypeInfo(index);
so->first = true;
so->probes = probes;
@@ -296,12 +284,6 @@ ivfflatbeginscan(Relation index, int nkeys, int norderbys)
so->normprocinfo = IvfflatOptionalProcInfo(index, IVFFLAT_NORM_PROC);
so->collation = index->rd_indcollation[0];
so->tmpCtx = AllocSetContextCreate(CurrentMemoryContext,
"Ivfflat scan temporary context",
ALLOCSET_DEFAULT_SIZES);
oldCtx = MemoryContextSwitchTo(so->tmpCtx);
/* Create tuple description for sorting */
so->tupdesc = CreateTemplateTupleDesc(2);
TupleDescInitEntry(so->tupdesc, (AttrNumber) 1, "distance", FLOAT8OID, -1, 0);
@@ -322,11 +304,8 @@ ivfflatbeginscan(Relation index, int nkeys, int norderbys)
so->bas = GetAccessStrategy(BAS_BULKREAD);
so->listQueue = pairingheap_allocate(CompareLists, scan);
so->listPages = palloc(maxProbes * sizeof(BlockNumber));
so->listIndex = 0;
so->lists = palloc(maxProbes * sizeof(IvfflatScanList));
MemoryContextSwitchTo(oldCtx);
so->startPages = palloc(maxProbes * sizeof(BlockNumber));
so->currentIndex = 0;
scan->opaque = so;
@@ -343,7 +322,7 @@ ivfflatrescan(IndexScanDesc scan, ScanKey keys, int nkeys, ScanKey orderbys, int
so->first = true;
pairingheap_reset(so->listQueue);
so->listIndex = 0;
so->currentIndex = 0;
if (keys && scan->numberOfKeys > 0)
memmove(scan->keyData, keys, scan->numberOfKeys * sizeof(ScanKeyData));
@@ -389,11 +368,13 @@ ivfflatgettuple(IndexScanDesc scan, ScanDirection dir)
IvfflatBench("GetScanItems", GetScanItems(scan, value));
so->first = false;
so->value = value;
/* TODO clean up if we allocated a new value */
}
while (!tuplesort_gettupleslot(so->sortstate, true, false, so->mslot, NULL))
{
if (so->listIndex == so->maxProbes)
if (so->currentIndex == so->maxProbes)
return false;
IvfflatBench("GetScanItems", GetScanItems(scan, so->value));
@@ -415,10 +396,13 @@ ivfflatendscan(IndexScanDesc scan)
{
IvfflatScanOpaque so = (IvfflatScanOpaque) scan->opaque;
/* Free any temporary files */
pairingheap_free(so->listQueue);
pfree(so->startPages);
tuplesort_end(so->sortstate);
FreeAccessStrategy(so->bas);
FreeTupleDesc(so->tupdesc);
MemoryContextDelete(so->tmpCtx);
/* TODO Free vslot and mslot without freeing TupleDesc */
pfree(so);
scan->opaque = NULL;

View File

@@ -26,7 +26,7 @@ ivfflatbulkdelete(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
Page cpage;
OffsetNumber coffno;
OffsetNumber cmaxoffno;
BlockNumber listPages[MaxOffsetNumber];
BlockNumber startPages[MaxOffsetNumber];
ListInfo listInfo;
cbuf = ReadBuffer(index, blkno);
@@ -40,7 +40,7 @@ ivfflatbulkdelete(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
{
IvfflatList list = (IvfflatList) PageGetItem(cpage, PageGetItemId(cpage, coffno));
listPages[coffno - FirstOffsetNumber] = list->startPage;
startPages[coffno - FirstOffsetNumber] = list->startPage;
}
listInfo.blkno = blkno;
@@ -50,7 +50,7 @@ ivfflatbulkdelete(IndexVacuumInfo *info, IndexBulkDeleteResult *stats,
for (coffno = FirstOffsetNumber; coffno <= cmaxoffno; coffno = OffsetNumberNext(coffno))
{
BlockNumber searchPage = listPages[coffno - FirstOffsetNumber];
BlockNumber searchPage = startPages[coffno - FirstOffsetNumber];
BlockNumber insertPage = InvalidBlockNumber;
/* Iterate over entry pages */

View File

@@ -155,6 +155,24 @@ CheckStateArray(ArrayType *statearray, const char *caller)
return (float8 *) ARR_DATA_PTR(statearray);
}
#if PG_VERSION_NUM < 120003
static pg_noinline void
float_overflow_error(void)
{
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("value out of range: overflow")));
}
static pg_noinline void
float_underflow_error(void)
{
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
errmsg("value out of range: underflow")));
}
#endif
/*
* Convert textual representation to internal representation
*/

View File

@@ -140,34 +140,6 @@ SELECT '{1e-8,-1e-8}'::real[]::halfvec;
[0,-0]
(1 row)
SELECT '[1,2,3]'::intvec::integer[];
int4
---------
{1,2,3}
(1 row)
SELECT '{1,2,3}'::integer[]::intvec;
intvec
---------
[1,2,3]
(1 row)
SELECT '{1,2,3}'::integer[]::intvec(3);
intvec
---------
[1,2,3]
(1 row)
SELECT '{1,2,3}'::integer[]::intvec(2);
ERROR: expected 2 dimensions, not 3
SELECT '{127,-128}'::integer[]::intvec;
intvec
------------
[127,-128]
(1 row)
SELECT '{128,-129}'::integer[]::intvec;
ERROR: value "128" is out of range for type intvec
SELECT '[0,1.5,0,3.5,0]'::vector::sparsevec;
sparsevec
-----------------

View File

@@ -30,23 +30,6 @@ SELECT * FROM t2 ORDER BY val;
(4 rows)
DROP TABLE t;
DROP TABLE t2;
-- intvec
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE TABLE t2 (val intvec(3));
\copy t TO 'results/intvec.bin' WITH (FORMAT binary)
\copy t2 FROM 'results/intvec.bin' WITH (FORMAT binary)
SELECT * FROM t2 ORDER BY val;
val
---------
[0,0,0]
[1,1,1]
[1,2,3]
(4 rows)
DROP TABLE t;
DROP TABLE t2;
-- sparsevec

View File

@@ -1,102 +0,0 @@
SET enable_seqscan = off;
-- L2
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_l2_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
[1,2,4]
[1,1,1]
[0,0,0]
(4 rows)
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <-> (SELECT NULL::intvec)) t2;
count
-------
4
(1 row)
SELECT COUNT(*) FROM t;
count
-------
5
(1 row)
TRUNCATE t;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
-----
(0 rows)
DROP TABLE t;
-- inner product
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_ip_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <#> '[3,3,3]';
val
---------
[1,2,4]
[1,2,3]
[1,1,1]
[0,0,0]
(4 rows)
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <#> (SELECT NULL::intvec)) t2;
count
-------
4
(1 row)
DROP TABLE t;
-- cosine
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_cosine_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <=> '[3,3,3]';
val
---------
[1,1,1]
[1,2,3]
[1,2,4]
(3 rows)
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <=> '[0,0,0]') t2;
count
-------
3
(1 row)
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <=> (SELECT NULL::intvec)) t2;
count
-------
3
(1 row)
DROP TABLE t;
-- L1
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_l1_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <+> '[3,3,3]';
val
---------
[1,2,3]
[1,2,4]
[1,1,1]
[0,0,0]
(4 rows)
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <+> (SELECT NULL::intvec)) t2;
count
-------
4
(1 row)
DROP TABLE t;

View File

@@ -99,32 +99,6 @@ SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <+> (SELECT NULL::vector)) t2
4
(1 row)
DROP TABLE t;
-- iterative
CREATE TABLE t (val vector(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val vector_l2_ops);
SET hnsw.iterative_search = strict_order;
SET hnsw.ef_search = 1;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
[1,1,1]
[0,0,0]
(3 rows)
SET hnsw.iterative_search = relaxed_order;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
[1,1,1]
[0,0,0]
(3 rows)
RESET hnsw.iterative_search;
RESET hnsw.ef_search;
DROP TABLE t;
-- unlogged
CREATE UNLOGGED TABLE t (val vector(3));
@@ -165,21 +139,4 @@ SET hnsw.ef_search = 0;
ERROR: 0 is outside the valid range for parameter "hnsw.ef_search" (1 .. 1000)
SET hnsw.ef_search = 1001;
ERROR: 1001 is outside the valid range for parameter "hnsw.ef_search" (1 .. 1000)
SHOW hnsw.iterative_search;
hnsw.iterative_search
-----------------------
off
(1 row)
SET hnsw.iterative_search = on;
ERROR: invalid value for parameter "hnsw.iterative_search": "on"
HINT: Available values: off, relaxed_order, strict_order.
SHOW hnsw.max_search_tuples;
hnsw.max_search_tuples
------------------------
-1
(1 row)
SET hnsw.max_search_tuples = -2;
ERROR: -2 is outside the valid range for parameter "hnsw.max_search_tuples" (-1 .. 2147483647)
DROP TABLE t;

View File

@@ -1,328 +0,0 @@
SELECT '[1,2,3]'::intvec;
intvec
---------
[1,2,3]
(1 row)
SELECT '[-1,-2,-3]'::intvec;
intvec
------------
[-1,-2,-3]
(1 row)
SELECT ' [ 1, 2 , 3 ] '::intvec;
intvec
---------
[1,2,3]
(1 row)
SELECT '[1.23456]'::intvec;
ERROR: invalid input syntax for type intvec: "[1.23456]"
LINE 1: SELECT '[1.23456]'::intvec;
^
SELECT '[hello,1]'::intvec;
ERROR: invalid input syntax for type intvec: "[hello,1]"
LINE 1: SELECT '[hello,1]'::intvec;
^
SELECT '[127,-128]'::intvec;
intvec
------------
[127,-128]
(1 row)
SELECT '[128,-129]'::intvec;
ERROR: "128" is out of range for type intvec
LINE 1: SELECT '[128,-129]'::intvec;
^
SELECT '[1,2,3'::intvec;
ERROR: invalid input syntax for type intvec: "[1,2,3"
LINE 1: SELECT '[1,2,3'::intvec;
^
SELECT '[1,2,3]9'::intvec;
ERROR: invalid input syntax for type intvec: "[1,2,3]9"
LINE 1: SELECT '[1,2,3]9'::intvec;
^
DETAIL: Junk after closing right brace.
SELECT '1,2,3'::intvec;
ERROR: invalid input syntax for type intvec: "1,2,3"
LINE 1: SELECT '1,2,3'::intvec;
^
DETAIL: Vector contents must start with "[".
SELECT ''::intvec;
ERROR: invalid input syntax for type intvec: ""
LINE 1: SELECT ''::intvec;
^
DETAIL: Vector contents must start with "[".
SELECT '['::intvec;
ERROR: invalid input syntax for type intvec: "["
LINE 1: SELECT '['::intvec;
^
SELECT '[,'::intvec;
ERROR: invalid input syntax for type intvec: "[,"
LINE 1: SELECT '[,'::intvec;
^
SELECT '[]'::intvec;
ERROR: intvec must have at least 1 dimension
LINE 1: SELECT '[]'::intvec;
^
SELECT '[1,]'::intvec;
ERROR: invalid input syntax for type intvec: "[1,]"
LINE 1: SELECT '[1,]'::intvec;
^
SELECT '[1a]'::intvec;
ERROR: invalid input syntax for type intvec: "[1a]"
LINE 1: SELECT '[1a]'::intvec;
^
SELECT '[1,,3]'::intvec;
ERROR: invalid input syntax for type intvec: "[1,,3]"
LINE 1: SELECT '[1,,3]'::intvec;
^
SELECT '[1, ,3]'::intvec;
ERROR: invalid input syntax for type intvec: "[1, ,3]"
LINE 1: SELECT '[1, ,3]'::intvec;
^
SELECT '[1,2,3]'::intvec(3);
intvec
---------
[1,2,3]
(1 row)
SELECT '[1,2,3]'::intvec(2);
ERROR: expected 2 dimensions, not 3
SELECT '[1,2,3]'::intvec(3, 2);
ERROR: invalid type modifier
LINE 1: SELECT '[1,2,3]'::intvec(3, 2);
^
SELECT '[1,2,3]'::intvec('a');
ERROR: invalid input syntax for type integer: "a"
LINE 1: SELECT '[1,2,3]'::intvec('a');
^
SELECT '[1,2,3]'::intvec(0);
ERROR: dimensions for type intvec must be at least 1
LINE 1: SELECT '[1,2,3]'::intvec(0);
^
SELECT '[1,2,3]'::intvec(16001);
ERROR: dimensions for type intvec cannot exceed 16000
LINE 1: SELECT '[1,2,3]'::intvec(16001);
^
SELECT unnest('{"[1,2,3]", "[4,5,6]"}'::intvec[]);
unnest
---------
[1,2,3]
[4,5,6]
(2 rows)
SELECT '{"[1,2,3]"}'::intvec(2)[];
ERROR: expected 2 dimensions, not 3
SELECT '[1,2,3]'::intvec < '[1,2,3]';
?column?
----------
f
(1 row)
SELECT '[1,2,3]'::intvec < '[1,2]';
?column?
----------
f
(1 row)
SELECT '[1,2,3]'::intvec <= '[1,2,3]';
?column?
----------
t
(1 row)
SELECT '[1,2,3]'::intvec <= '[1,2]';
?column?
----------
f
(1 row)
SELECT '[1,2,3]'::intvec = '[1,2,3]';
?column?
----------
t
(1 row)
SELECT '[1,2,3]'::intvec = '[1,2]';
?column?
----------
f
(1 row)
SELECT '[1,2,3]'::intvec != '[1,2,3]';
?column?
----------
f
(1 row)
SELECT '[1,2,3]'::intvec != '[1,2]';
?column?
----------
t
(1 row)
SELECT '[1,2,3]'::intvec >= '[1,2,3]';
?column?
----------
t
(1 row)
SELECT '[1,2,3]'::intvec >= '[1,2]';
?column?
----------
t
(1 row)
SELECT '[1,2,3]'::intvec > '[1,2,3]';
?column?
----------
f
(1 row)
SELECT '[1,2,3]'::intvec > '[1,2]';
?column?
----------
t
(1 row)
SELECT intvec_cmp('[1,2,3]', '[1,2,3]');
intvec_cmp
------------
0
(1 row)
SELECT intvec_cmp('[1,2,3]', '[0,0,0]');
intvec_cmp
------------
1
(1 row)
SELECT intvec_cmp('[0,0,0]', '[1,2,3]');
intvec_cmp
------------
-1
(1 row)
SELECT intvec_cmp('[1,2]', '[1,2,3]');
intvec_cmp
------------
-1
(1 row)
SELECT intvec_cmp('[1,2,3]', '[1,2]');
intvec_cmp
------------
1
(1 row)
SELECT intvec_cmp('[1,2]', '[2,3,4]');
intvec_cmp
------------
-1
(1 row)
SELECT intvec_cmp('[2,3]', '[1,2,3]');
intvec_cmp
------------
1
(1 row)
SELECT vector_dims('[1,2,3]'::intvec);
vector_dims
-------------
3
(1 row)
SELECT l2_distance('[0,0]'::intvec, '[3,4]');
l2_distance
-------------
5
(1 row)
SELECT l2_distance('[0,0]'::intvec, '[0,1]');
l2_distance
-------------
1
(1 row)
SELECT l2_distance('[1,2]'::intvec, '[3]');
ERROR: different intvec dimensions 2 and 1
SELECT '[0,0]'::intvec <-> '[3,4]';
?column?
----------
5
(1 row)
SELECT inner_product('[1,2]'::intvec, '[3,4]');
inner_product
---------------
11
(1 row)
SELECT inner_product('[1,2]'::intvec, '[3]');
ERROR: different intvec dimensions 2 and 1
SELECT inner_product('[127]'::intvec, '[127]');
inner_product
---------------
16129
(1 row)
SELECT '[1,2]'::intvec <#> '[3,4]';
?column?
----------
-11
(1 row)
SELECT cosine_distance('[1,2]'::intvec, '[2,4]');
cosine_distance
-----------------
0
(1 row)
SELECT cosine_distance('[1,2]'::intvec, '[0,0]');
cosine_distance
-----------------
NaN
(1 row)
SELECT cosine_distance('[1,1]'::intvec, '[1,1]');
cosine_distance
-----------------
0
(1 row)
SELECT cosine_distance('[1,0]'::intvec, '[0,2]');
cosine_distance
-----------------
1
(1 row)
SELECT cosine_distance('[1,1]'::intvec, '[-1,-1]');
cosine_distance
-----------------
2
(1 row)
SELECT cosine_distance('[1,2]'::intvec, '[3]');
ERROR: different intvec dimensions 2 and 1
SELECT '[1,2]'::intvec <=> '[2,4]';
?column?
----------
0
(1 row)
SELECT l1_distance('[0,0]'::intvec, '[3,4]');
l1_distance
-------------
7
(1 row)
SELECT l1_distance('[0,0]'::intvec, '[0,1]');
l1_distance
-------------
1
(1 row)
SELECT l1_distance('[1,2]'::intvec, '[3]');
ERROR: different intvec dimensions 2 and 1

View File

@@ -81,44 +81,6 @@ SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <=> (SELECT NULL::vector)) t2
3
(1 row)
DROP TABLE t;
-- iterative
CREATE TABLE t (val vector(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING ivfflat (val vector_l2_ops) WITH (lists = 3);
SET ivfflat.iterative_search = relaxed_order;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
[1,1,1]
[0,0,0]
(3 rows)
SET ivfflat.max_probes = 0;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
(1 row)
SET ivfflat.max_probes = 1;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
(1 row)
SET ivfflat.max_probes = 2;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
val
---------
[1,2,3]
[1,1,1]
(2 rows)
RESET ivfflat.iterative_search;
RESET ivfflat.max_probes;
DROP TABLE t;
-- unlogged
CREATE UNLOGGED TABLE t (val vector(3));
@@ -147,27 +109,4 @@ SHOW ivfflat.probes;
1
(1 row)
SET ivfflat.probes = 0;
ERROR: 0 is outside the valid range for parameter "ivfflat.probes" (1 .. 32768)
SET ivfflat.probes = 32769;
ERROR: 32769 is outside the valid range for parameter "ivfflat.probes" (1 .. 32768)
SHOW ivfflat.iterative_search;
ivfflat.iterative_search
--------------------------
off
(1 row)
SET ivfflat.iterative_search = on;
ERROR: invalid value for parameter "ivfflat.iterative_search": "on"
HINT: Available values: off, relaxed_order.
SHOW ivfflat.max_probes;
ivfflat.max_probes
--------------------
-1
(1 row)
SET ivfflat.max_probes = -2;
ERROR: -2 is outside the valid range for parameter "ivfflat.max_probes" (-1 .. 32768)
SET ivfflat.max_probes = 32769;
ERROR: 32769 is outside the valid range for parameter "ivfflat.max_probes" (-1 .. 32768)
DROP TABLE t;

View File

@@ -38,14 +38,6 @@ SELECT '{1,2,3}'::real[]::halfvec(2);
SELECT '{65520,-65520}'::real[]::halfvec;
SELECT '{1e-8,-1e-8}'::real[]::halfvec;
SELECT '[1,2,3]'::intvec::integer[];
SELECT '{1,2,3}'::integer[]::intvec;
SELECT '{1,2,3}'::integer[]::intvec(3);
SELECT '{1,2,3}'::integer[]::intvec(2);
SELECT '{127,-128}'::integer[]::intvec;
SELECT '{128,-129}'::integer[]::intvec;
SELECT '[0,1.5,0,3.5,0]'::vector::sparsevec;
SELECT '[0,1.5,0,3.5,0]'::vector::sparsevec(5);
SELECT '[0,1.5,0,3.5,0]'::vector::sparsevec(4);

View File

@@ -28,21 +28,6 @@ SELECT * FROM t2 ORDER BY val;
DROP TABLE t;
DROP TABLE t2;
-- intvec
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE TABLE t2 (val intvec(3));
\copy t TO 'results/intvec.bin' WITH (FORMAT binary)
\copy t2 FROM 'results/intvec.bin' WITH (FORMAT binary)
SELECT * FROM t2 ORDER BY val;
DROP TABLE t;
DROP TABLE t2;
-- sparsevec
CREATE TABLE t (val sparsevec(3));

View File

@@ -1,58 +0,0 @@
SET enable_seqscan = off;
-- L2
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_l2_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <-> (SELECT NULL::intvec)) t2;
SELECT COUNT(*) FROM t;
TRUNCATE t;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
DROP TABLE t;
-- inner product
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_ip_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <#> '[3,3,3]';
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <#> (SELECT NULL::intvec)) t2;
DROP TABLE t;
-- cosine
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_cosine_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <=> '[3,3,3]';
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <=> '[0,0,0]') t2;
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <=> (SELECT NULL::intvec)) t2;
DROP TABLE t;
-- L1
CREATE TABLE t (val intvec(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val intvec_l1_ops);
INSERT INTO t (val) VALUES ('[1,2,4]');
SELECT * FROM t ORDER BY val <+> '[3,3,3]';
SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <+> (SELECT NULL::intvec)) t2;
DROP TABLE t;

View File

@@ -57,23 +57,6 @@ SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <+> (SELECT NULL::vector)) t2
DROP TABLE t;
-- iterative
CREATE TABLE t (val vector(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING hnsw (val vector_l2_ops);
SET hnsw.iterative_search = strict_order;
SET hnsw.ef_search = 1;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
SET hnsw.iterative_search = relaxed_order;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
RESET hnsw.iterative_search;
RESET hnsw.ef_search;
DROP TABLE t;
-- unlogged
CREATE UNLOGGED TABLE t (val vector(3));
@@ -98,12 +81,4 @@ SHOW hnsw.ef_search;
SET hnsw.ef_search = 0;
SET hnsw.ef_search = 1001;
SHOW hnsw.iterative_search;
SET hnsw.iterative_search = on;
SHOW hnsw.max_search_tuples;
SET hnsw.max_search_tuples = -2;
DROP TABLE t;

View File

@@ -1,73 +0,0 @@
SELECT '[1,2,3]'::intvec;
SELECT '[-1,-2,-3]'::intvec;
SELECT ' [ 1, 2 , 3 ] '::intvec;
SELECT '[1.23456]'::intvec;
SELECT '[hello,1]'::intvec;
SELECT '[127,-128]'::intvec;
SELECT '[128,-129]'::intvec;
SELECT '[1,2,3'::intvec;
SELECT '[1,2,3]9'::intvec;
SELECT '1,2,3'::intvec;
SELECT ''::intvec;
SELECT '['::intvec;
SELECT '[,'::intvec;
SELECT '[]'::intvec;
SELECT '[1,]'::intvec;
SELECT '[1a]'::intvec;
SELECT '[1,,3]'::intvec;
SELECT '[1, ,3]'::intvec;
SELECT '[1,2,3]'::intvec(3);
SELECT '[1,2,3]'::intvec(2);
SELECT '[1,2,3]'::intvec(3, 2);
SELECT '[1,2,3]'::intvec('a');
SELECT '[1,2,3]'::intvec(0);
SELECT '[1,2,3]'::intvec(16001);
SELECT unnest('{"[1,2,3]", "[4,5,6]"}'::intvec[]);
SELECT '{"[1,2,3]"}'::intvec(2)[];
SELECT '[1,2,3]'::intvec < '[1,2,3]';
SELECT '[1,2,3]'::intvec < '[1,2]';
SELECT '[1,2,3]'::intvec <= '[1,2,3]';
SELECT '[1,2,3]'::intvec <= '[1,2]';
SELECT '[1,2,3]'::intvec = '[1,2,3]';
SELECT '[1,2,3]'::intvec = '[1,2]';
SELECT '[1,2,3]'::intvec != '[1,2,3]';
SELECT '[1,2,3]'::intvec != '[1,2]';
SELECT '[1,2,3]'::intvec >= '[1,2,3]';
SELECT '[1,2,3]'::intvec >= '[1,2]';
SELECT '[1,2,3]'::intvec > '[1,2,3]';
SELECT '[1,2,3]'::intvec > '[1,2]';
SELECT intvec_cmp('[1,2,3]', '[1,2,3]');
SELECT intvec_cmp('[1,2,3]', '[0,0,0]');
SELECT intvec_cmp('[0,0,0]', '[1,2,3]');
SELECT intvec_cmp('[1,2]', '[1,2,3]');
SELECT intvec_cmp('[1,2,3]', '[1,2]');
SELECT intvec_cmp('[1,2]', '[2,3,4]');
SELECT intvec_cmp('[2,3]', '[1,2,3]');
SELECT vector_dims('[1,2,3]'::intvec);
SELECT l2_distance('[0,0]'::intvec, '[3,4]');
SELECT l2_distance('[0,0]'::intvec, '[0,1]');
SELECT l2_distance('[1,2]'::intvec, '[3]');
SELECT '[0,0]'::intvec <-> '[3,4]';
SELECT inner_product('[1,2]'::intvec, '[3,4]');
SELECT inner_product('[1,2]'::intvec, '[3]');
SELECT inner_product('[127]'::intvec, '[127]');
SELECT '[1,2]'::intvec <#> '[3,4]';
SELECT cosine_distance('[1,2]'::intvec, '[2,4]');
SELECT cosine_distance('[1,2]'::intvec, '[0,0]');
SELECT cosine_distance('[1,1]'::intvec, '[1,1]');
SELECT cosine_distance('[1,0]'::intvec, '[0,2]');
SELECT cosine_distance('[1,1]'::intvec, '[-1,-1]');
SELECT cosine_distance('[1,2]'::intvec, '[3]');
SELECT '[1,2]'::intvec <=> '[2,4]';
SELECT l1_distance('[0,0]'::intvec, '[3,4]');
SELECT l1_distance('[0,0]'::intvec, '[0,1]');
SELECT l1_distance('[1,2]'::intvec, '[3]');

View File

@@ -44,28 +44,6 @@ SELECT COUNT(*) FROM (SELECT * FROM t ORDER BY val <=> (SELECT NULL::vector)) t2
DROP TABLE t;
-- iterative
CREATE TABLE t (val vector(3));
INSERT INTO t (val) VALUES ('[0,0,0]'), ('[1,2,3]'), ('[1,1,1]'), (NULL);
CREATE INDEX ON t USING ivfflat (val vector_l2_ops) WITH (lists = 3);
SET ivfflat.iterative_search = relaxed_order;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
SET ivfflat.max_probes = 0;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
SET ivfflat.max_probes = 1;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
SET ivfflat.max_probes = 2;
SELECT * FROM t ORDER BY val <-> '[3,3,3]';
RESET ivfflat.iterative_search;
RESET ivfflat.max_probes;
DROP TABLE t;
-- unlogged
CREATE UNLOGGED TABLE t (val vector(3));
@@ -84,16 +62,4 @@ CREATE INDEX ON t USING ivfflat (val vector_l2_ops) WITH (lists = 32769);
SHOW ivfflat.probes;
SET ivfflat.probes = 0;
SET ivfflat.probes = 32769;
SHOW ivfflat.iterative_search;
SET ivfflat.iterative_search = on;
SHOW ivfflat.max_probes;
SET ivfflat.max_probes = -2;
SET ivfflat.max_probes = 32769;
DROP TABLE t;

View File

@@ -6,7 +6,13 @@ use Test::More;
my $dim = 3;
my $array_sql = join(",", ('random()') x $dim);
my @r = ();
for (1 .. $dim)
{
my $v = int(rand(1000)) + 1;
push(@r, "i % $v");
}
my $array_sql = join(", ", @r);
# Initialize node
my $node = PostgreSQL::Test::Cluster->new('node');
@@ -17,20 +23,19 @@ $node->start;
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v vector($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 100000) i;"
"INSERT INTO tst SELECT i % 10, ARRAY[$array_sql] FROM generate_series(1, 100000) i;"
);
$node->safe_psql("postgres", "CREATE INDEX ON tst USING ivfflat (v vector_l2_ops);");
# Get size
my $size = $node->safe_psql("postgres", "SELECT pg_total_relation_size('tst_v_idx');");
# Store values
$node->safe_psql("postgres", "CREATE TABLE tmp AS SELECT * FROM tst;");
# Delete all, vacuum, and insert same data
$node->safe_psql("postgres", "DELETE FROM tst;");
$node->safe_psql("postgres", "VACUUM tst;");
$node->safe_psql("postgres", "INSERT INTO tst SELECT * FROM tmp;");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i % 10, ARRAY[$array_sql] FROM generate_series(1, 100000) i;"
);
# Check size
my $new_size = $node->safe_psql("postgres", "SELECT pg_total_relation_size('tst_v_idx');");

View File

@@ -94,7 +94,8 @@ like($explain, qr/Seq Scan/);
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE v <-> '$query' < 1 ORDER BY v <-> '$query';
));
like($explain, qr/Seq Scan/);
# TODO Do not use index
like($explain, qr/Index Scan using idx/);
# Test attribute index
$node->safe_psql("postgres", "CREATE INDEX attribute_idx ON tst (c);");
@@ -109,6 +110,7 @@ $node->safe_psql("postgres", "CREATE INDEX partial_idx ON tst USING ivfflat (v v
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE c = $c ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using partial_idx/);
# TODO Use partial index
like($explain, qr/Index Scan using idx/);
done_testing();

View File

@@ -18,13 +18,9 @@ $node->start;
# Create table and index
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v vector($dim), c int4, t text);");
$node->safe_psql("postgres", "CREATE TABLE cat (i int4 PRIMARY KEY, t text, b boolean);");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql], i % $nc, 'test ' || i FROM generate_series(1, 10000) i;"
);
$node->safe_psql("postgres",
"INSERT INTO cat SELECT i, 'cat ' || i, i % 5 = 0 FROM generate_series(1, $nc) i;"
);
$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v vector_l2_ops);");
$node->safe_psql("postgres", "ANALYZE tst;");
@@ -41,7 +37,8 @@ my $c = int(rand() * $nc);
my $explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE c = $c ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Seq Scan/);
# TODO Do not use index
like($explain, qr/Index Scan using idx/);
# Test attribute filtering with few rows removed
$explain = $node->safe_psql("postgres", qq(
@@ -59,7 +56,8 @@ like($explain, qr/Index Scan using idx/);
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE c < 1 ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Seq Scan/);
# TODO Do not use index
like($explain, qr/Index Scan using idx/);
# Test attribute filtering with few rows removed like
$explain = $node->safe_psql("postgres", qq(
@@ -98,25 +96,13 @@ $explain = $node->safe_psql("postgres", qq(
));
like($explain, qr/Seq Scan/);
# Test join
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT cat.t FROM cat INNER JOIN tst ON cat.i = tst.c ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using idx/);
# Test join with attribute filtering
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT cat.t FROM cat INNER JOIN tst ON cat.i = tst.c WHERE cat.b = 't' ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using idx/);
# Test attribute index
$node->safe_psql("postgres", "CREATE INDEX attribute_idx ON tst (c);");
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE c = $c ORDER BY v <-> '$query' LIMIT $limit;
));
# Use attribute index
like($explain, qr/Bitmap Index Scan on attribute_idx/);
# TODO Use attribute index
like($explain, qr/Index Scan using idx/);
# Test partial index
$node->safe_psql("postgres", "CREATE INDEX partial_idx ON tst USING hnsw (v vector_l2_ops) WHERE (c = $c);");

View File

@@ -40,10 +40,6 @@ for (1 .. 50)
$actual = $node->safe_psql("postgres", "SELECT halfvec_cmp(v::halfvec, '$query'::real[]::halfvec) FROM tst");
is($expected, $actual);
# Test intvec
$actual = $node->safe_psql("postgres", "SELECT intvec_cmp(v::integer[]::intvec, '$query'::integer[]::intvec) FROM tst");
is($expected, $actual);
# Test sparsevec
$actual = $node->safe_psql("postgres", "SELECT sparsevec_cmp(v::vector::sparsevec, '$query'::real[]::vector::sparsevec) FROM tst");
is($expected, $actual);

View File

@@ -45,10 +45,6 @@ for my $function (@functions)
my $actual = $node->safe_psql("postgres", "SELECT $function(v::halfvec, '$query'::vector::halfvec) FROM tst");
is($expected, $actual, "halfvec $function");
# Test intvec
$actual = $node->safe_psql("postgres", "SELECT $function(v::real[]::integer[]::intvec, '$query'::vector::real[]::integer[]::intvec) FROM tst");
is($expected, $actual, "intvec $function");
# Test sparsevec
$actual = $node->safe_psql("postgres", "SELECT $function(v::sparsevec, '$query'::vector::sparsevec) FROM tst");
is($expected, $actual, "sparsevec $function");

View File

@@ -12,8 +12,8 @@ $node->start;
# Create extension
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
my @types = ("vector", "halfvec", "intvec", "sparsevec");
my @inputs = ("[1.23,4.56,7.89]", "[1.23,4.56,7.89]", "[1,2,3]", "{1:1.23,2:4.56,3:7.89}/3");
my @types = ("vector", "halfvec", "sparsevec");
my @inputs = ("[1.23,4.56,7.89]", "[1.23,4.56,7.89]", "{1:1.23,2:4.56,3:7.89}/3");
my @subs = (" ", " ", ",", ":", "-", "1", "9", "\0", "2147483648", "-2147483649");
for my $i (0 .. $#types)

View File

@@ -1,60 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my @dims = (384, 1536);
my $limit = 10;
# Initialize node
my $node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
for my $dim (@dims)
{
my $array_sql = join(",", ('random()') x $dim);
# Create table and index
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v vector($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 2000) i;"
);
$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v vector_l2_ops);");
$node->safe_psql("postgres", "ANALYZE tst;");
# Generate query
my @r = ();
for (1 .. $dim)
{
push(@r, rand());
}
my $query = "[" . join(",", @r) . "]";
my $explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using idx/);
# 3x the rows are needed for distance filters
# since the planner uses DEFAULT_INEQ_SEL for the selectivity (should be 1)
# Recreate index for performance
$node->safe_psql("postgres", "DROP INDEX idx;");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(2001, 6000) i;"
);
$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v vector_l2_ops);");
$node->safe_psql("postgres", "ANALYZE tst;");
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE v <-> '$query' < 1 ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using idx/);
$node->safe_psql("postgres", "DROP TABLE tst;");
}
done_testing();

View File

@@ -14,7 +14,7 @@ $node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4 PRIMARY KEY, v vector($dim));");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v vector($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 100000) i;"
);
@@ -23,32 +23,9 @@ $node->safe_psql("postgres", "CREATE INDEX ON tst USING ivfflat (v vector_l2_ops
my $count = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET ivfflat.probes = 10;
SET ivfflat.iterative_search = relaxed_order;
SET ivfflat.streaming = on;
SELECT COUNT(*) FROM (SELECT v FROM tst WHERE i % 10000 = 0 ORDER BY v <-> (SELECT v FROM tst LIMIT 1) LIMIT 11) t;
));
is($count, 10);
foreach ((30, 50, 70))
{
my $max_probes = $_;
my $expected = $max_probes / 10;
my $sum = 0;
for my $i (1 .. 20)
{
$count = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET ivfflat.probes = 10;
SET ivfflat.iterative_search = relaxed_order;
SET ivfflat.max_probes = $max_probes;
SELECT COUNT(*) FROM (SELECT v FROM tst WHERE i % 10000 = 0 ORDER BY v <-> (SELECT v FROM tst WHERE i = $i) LIMIT 11) t;
));
$sum += $count;
}
my $avg = $sum / 20;
cmp_ok($avg, '>', $expected - 2);
cmp_ok($avg, '<', $expected + 2);
}
done_testing();

View File

@@ -1,50 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my @dims = (384, 1536);
my $limit = 10;
# Initialize node
my $node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
for my $dim (@dims)
{
my $array_sql = join(",", ('random()') x $dim);
# Create table and index
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v vector($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 5000) i;"
);
$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING ivfflat (v vector_l2_ops) WITH (lists = 5);");
$node->safe_psql("postgres", "ANALYZE tst;");
# Generate query
my @r = ();
for (1 .. $dim)
{
push(@r, rand());
}
my $query = "[" . join(",", @r) . "]";
my $explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using idx/);
$explain = $node->safe_psql("postgres", qq(
EXPLAIN ANALYZE SELECT i FROM tst WHERE v <-> '$query' < 1 ORDER BY v <-> '$query' LIMIT $limit;
));
like($explain, qr/Index Scan using idx/);
$node->safe_psql("postgres", "DROP TABLE tst;");
}
done_testing();

View File

@@ -19,7 +19,7 @@ sub test_recall
my $explain = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET ivfflat.probes = $probes;
SET ivfflat.iterative_search = relaxed_order;
SET ivfflat.streaming = on;
EXPLAIN ANALYZE SELECT i FROM tst WHERE i % $c = 0 ORDER BY v $operator '$queries[0]' LIMIT $limit;
));
like($explain, qr/Index Scan using idx on tst/);
@@ -29,7 +29,7 @@ sub test_recall
my $actual = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET ivfflat.probes = $probes;
SET ivfflat.iterative_search = relaxed_order;
SET ivfflat.streaming = on;
SELECT i FROM tst WHERE i % $c = 0 ORDER BY v $operator '$queries[$i]' LIMIT $limit;
));
my @actual_ids = split("\n", $actual);
@@ -48,7 +48,7 @@ sub test_recall
$total += $limit;
}
cmp_ok($correct / $total, ">=", $min, "$operator $c");
cmp_ok($correct / $total, ">=", $min, $operator);
}
# Initialize node
@@ -103,7 +103,7 @@ for my $i (0 .. $#operators)
if ($c == 100)
{
test_recall($c, 1, 0.57, $operator);
test_recall($c, 1, 0.58, $operator);
test_recall($c, 10, 0.98, $operator);
}
else

View File

@@ -1,67 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my $dim = 3;
my $array_sql = join(",", ('random()') x $dim);
# Initialize node
my $node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4 PRIMARY KEY, v vector($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 100000) i;"
);
$node->safe_psql("postgres", qq(
SET maintenance_work_mem = '128MB';
SET max_parallel_maintenance_workers = 2;
CREATE INDEX ON tst USING hnsw (v vector_l2_ops)
));
my $count = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.iterative_search = relaxed_order;
SET work_mem = '8MB';
SELECT COUNT(*) FROM (SELECT v FROM tst WHERE i % 10000 = 0 ORDER BY v <-> (SELECT v FROM tst LIMIT 1) LIMIT 11) t;
));
is($count, 10);
foreach ((30000, 50000, 70000))
{
my $max_tuples = $_;
my $expected = $max_tuples / 10000;
my $sum = 0;
for my $i (1 .. 20)
{
$count = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.iterative_search = relaxed_order;
SET hnsw.max_search_tuples = $max_tuples;
SET work_mem = '8MB';
SELECT COUNT(*) FROM (SELECT v FROM tst WHERE i % 10000 = 0 ORDER BY v <-> (SELECT v FROM tst WHERE i = $i) LIMIT 11) t;
));
$sum += $count;
}
my $avg = $sum / 20;
cmp_ok($avg, '>', $expected - 2);
cmp_ok($avg, '<', $expected + 2);
}
my ($ret, $stdout, $stderr) = $node->psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.iterative_search = relaxed_order;
SET client_min_messages = debug1;
SET work_mem = '2MB';
SELECT COUNT(*) FROM (SELECT v FROM tst WHERE i % 10000 = 0 ORDER BY v <-> (SELECT v FROM tst LIMIT 1) LIMIT 11) t;
));
like($stderr, qr/hnsw index scan exceeded work_mem after \d+ tuples/);
done_testing();

View File

@@ -1,118 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my $node;
my @queries = ();
my @expected;
my $limit = 20;
my $dim = 3;
my $array_sql = join(",", ('random()') x $dim);
my @cs = (50, 500);
sub test_recall
{
my ($c, $ef_search, $min, $operator, $mode) = @_;
my $correct = 0;
my $total = 0;
my $explain = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.ef_search = $ef_search;
SET hnsw.iterative_search = $mode;
EXPLAIN ANALYZE SELECT i FROM tst WHERE i % $c = 0 ORDER BY v $operator '$queries[0]' LIMIT $limit;
));
like($explain, qr/Index Scan using idx on tst/);
for my $i (0 .. $#queries)
{
my $actual = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.ef_search = $ef_search;
SET hnsw.iterative_search = $mode;
SELECT i FROM tst WHERE i % $c = 0 ORDER BY v $operator '$queries[$i]' LIMIT $limit;
));
my @actual_ids = split("\n", $actual);
my @expected_ids = split("\n", $expected[$i]);
my %expected_set = map { $_ => 1 } @expected_ids;
foreach (@actual_ids)
{
if (exists($expected_set{$_}))
{
$correct++;
}
}
$total += $limit;
}
cmp_ok($correct / $total, ">=", $min, "$operator $mode $c");
}
# Initialize node
$node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v vector($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 50000) i;"
);
# Generate queries
for (1 .. 20)
{
my @r = ();
for (1 .. $dim)
{
push(@r, rand());
}
push(@queries, "[" . join(",", @r) . "]");
}
# Check each index type
my @operators = ("<->", "<=>");
my @opclasses = ("vector_l2_ops", "vector_cosine_ops");
for my $i (0 .. $#operators)
{
my $operator = $operators[$i];
my $opclass = $opclasses[$i];
$node->safe_psql("postgres", qq(
SET maintenance_work_mem = '128MB';
CREATE INDEX idx ON tst USING hnsw (v $opclass);
));
foreach (@cs)
{
my $c = $_;
# Get exact results
@expected = ();
foreach (@queries)
{
my $res = $node->safe_psql("postgres", qq(
SET enable_indexscan = off;
WITH top AS (
SELECT v $operator '$_' AS distance FROM tst WHERE i % $c = 0 ORDER BY distance LIMIT $limit
)
SELECT i FROM tst WHERE (v $operator '$_') <= (SELECT MAX(distance) FROM top)
));
push(@expected, $res);
}
test_recall($c, 40, 0.99, $operator, "strict_order");
test_recall($c, 40, 0.99, $operator, "relaxed_order");
}
$node->safe_psql("postgres", "DROP INDEX idx;");
}
done_testing();

View File

@@ -1,132 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my $node;
my @queries = ();
my @expected;
my $limit = 20;
my $dim = 10;
my $array_sql = join(",", ('(random() * 255)::int - 128') x $dim);
sub test_recall
{
my ($min, $operator) = @_;
my $correct = 0;
my $total = 0;
my $explain = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
EXPLAIN ANALYZE SELECT i FROM tst ORDER BY v $operator '$queries[0]' LIMIT $limit;
));
like($explain, qr/Index Scan/);
for my $i (0 .. $#queries)
{
my $actual = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SELECT i FROM tst ORDER BY v $operator '$queries[$i]' LIMIT $limit;
));
my @actual_ids = split("\n", $actual);
my %actual_set = map { $_ => 1 } @actual_ids;
my @expected_ids = split("\n", $expected[$i]);
foreach (@expected_ids)
{
if (exists($actual_set{$_}))
{
$correct++;
}
$total++;
}
}
cmp_ok($correct / $total, ">=", $min, $operator);
}
# Initialize node
$node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v intvec($dim));");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 10000) i;"
);
# Generate queries
for (1 .. 20)
{
my @r = ();
for (1 .. $dim)
{
push(@r, int(rand(256)) - 128);
}
push(@queries, "[" . join(",", @r) . "]");
}
# Check each index type
my @operators = ("<->", "<#>", "<=>", "<+>");
my @opclasses = ("intvec_l2_ops", "intvec_ip_ops", "intvec_cosine_ops", "intvec_l1_ops");
for my $i (0 .. $#operators)
{
my $operator = $operators[$i];
my $opclass = $opclasses[$i];
# Get exact results
@expected = ();
foreach (@queries)
{
my $res = $node->safe_psql("postgres", "SELECT i FROM tst ORDER BY v $operator '$_' LIMIT $limit;");
push(@expected, $res);
}
# Build index serially
$node->safe_psql("postgres", qq(
SET max_parallel_maintenance_workers = 0;
CREATE INDEX idx ON tst USING hnsw (v $opclass);
));
# Test approximate results
my $min = 0.98;
test_recall($min, $operator);
$node->safe_psql("postgres", "DROP INDEX idx;");
# Build index in parallel in memory
my ($ret, $stdout, $stderr) = $node->psql("postgres", qq(
SET client_min_messages = DEBUG;
SET min_parallel_table_scan_size = 1;
CREATE INDEX idx ON tst USING hnsw (v $opclass);
));
is($ret, 0, $stderr);
like($stderr, qr/using \d+ parallel workers/);
# Test approximate results
test_recall($min, $operator);
$node->safe_psql("postgres", "DROP INDEX idx;");
# Build index in parallel on disk
# Set parallel_workers on table to use workers with low maintenance_work_mem
($ret, $stdout, $stderr) = $node->psql("postgres", qq(
ALTER TABLE tst SET (parallel_workers = 2);
SET client_min_messages = DEBUG;
SET maintenance_work_mem = '4MB';
CREATE INDEX idx ON tst USING hnsw (v $opclass);
ALTER TABLE tst RESET (parallel_workers);
));
is($ret, 0, $stderr);
like($stderr, qr/using \d+ parallel workers/);
like($stderr, qr/hnsw graph no longer fits into maintenance_work_mem/);
$node->safe_psql("postgres", "DROP INDEX idx;");
}
done_testing();

View File

@@ -1,113 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my $node;
my @queries = ();
my @expected;
my $limit = 20;
my $dim = 10;
my $array_sql = join(",", ('(random() * 255)::int - 128') x $dim);
sub test_recall
{
my ($min, $operator) = @_;
my $correct = 0;
my $total = 0;
my $explain = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
EXPLAIN ANALYZE SELECT i FROM tst ORDER BY v $operator '$queries[0]' LIMIT $limit;
));
like($explain, qr/Index Scan/);
for my $i (0 .. $#queries)
{
my $actual = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SELECT i FROM tst ORDER BY v $operator '$queries[$i]' LIMIT $limit;
));
my @actual_ids = split("\n", $actual);
my %actual_set = map { $_ => 1 } @actual_ids;
my @expected_ids = split("\n", $expected[$i]);
foreach (@expected_ids)
{
if (exists($actual_set{$_}))
{
$correct++;
}
$total++;
}
}
cmp_ok($correct / $total, ">=", $min, $operator);
}
# Initialize node
$node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i serial, v intvec($dim));");
# Generate queries
for (1 .. 20)
{
my @r = ();
for (1 .. $dim)
{
push(@r, int(rand(256)) - 128);
}
push(@queries, "[" . join(",", @r) . "]");
}
# Check each index type
my @operators = ("<->", "<#>", "<=>", "<+>");
my @opclasses = ("intvec_l2_ops", "intvec_ip_ops", "intvec_cosine_ops", "intvec_l1_ops");
for my $i (0 .. $#operators)
{
my $operator = $operators[$i];
my $opclass = $opclasses[$i];
# Add index
$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v $opclass);");
# Use concurrent inserts
$node->pgbench(
"--no-vacuum --client=10 --transactions=1000",
0,
[qr{actually processed}],
[qr{^$}],
"concurrent INSERTs",
{
"040_hnsw_intvec_insert_recall_$opclass" => "INSERT INTO tst (v) VALUES (ARRAY[$array_sql]);"
}
);
# Get exact results
@expected = ();
foreach (@queries)
{
my $res = $node->safe_psql("postgres", qq(
SET enable_indexscan = off;
SELECT i FROM tst ORDER BY v $operator '$_' LIMIT $limit;
));
push(@expected, $res);
}
# Test approximate results
my $min = 0.98;
test_recall($min, $operator);
$node->safe_psql("postgres", "DROP INDEX idx;");
$node->safe_psql("postgres", "TRUNCATE tst;");
}
done_testing();

View File

@@ -1,101 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
my $node;
my @queries = ();
my @expected;
my $limit = 20;
my $dim = 10;
my $array_sql = join(",", ('(random() * 255)::int - 128') x $dim);
sub test_recall
{
my ($min, $ef_search, $test_name) = @_;
my $correct = 0;
my $total = 0;
my $explain = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.ef_search = $ef_search;
EXPLAIN ANALYZE SELECT i FROM tst ORDER BY v <-> '$queries[0]' LIMIT $limit;
));
like($explain, qr/Index Scan/);
for my $i (0 .. $#queries)
{
my $actual = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.ef_search = $ef_search;
SELECT i FROM tst ORDER BY v <-> '$queries[$i]' LIMIT $limit;
));
my @actual_ids = split("\n", $actual);
my %actual_set = map { $_ => 1 } @actual_ids;
my @expected_ids = split("\n", $expected[$i]);
foreach (@expected_ids)
{
if (exists($actual_set{$_}))
{
$correct++;
}
$total++;
}
}
cmp_ok($correct / $total, ">=", $min, $test_name);
}
# Initialize node
$node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (i int4, v intvec($dim));");
$node->safe_psql("postgres", "ALTER TABLE tst SET (autovacuum_enabled = false);");
$node->safe_psql("postgres",
"INSERT INTO tst SELECT i, ARRAY[$array_sql] FROM generate_series(1, 10000) i;"
);
# Add index
$node->safe_psql("postgres", "CREATE INDEX ON tst USING hnsw (v intvec_l2_ops) WITH (m = 4, ef_construction = 8);");
# Delete data
$node->safe_psql("postgres", "DELETE FROM tst WHERE i > 2500;");
# Generate queries
for (1 .. 20)
{
my @r = ();
for (1 .. $dim)
{
push(@r, int(rand(256)) - 128);
}
push(@queries, "[" . join(",", @r) . "]");
}
# Get exact results
@expected = ();
foreach (@queries)
{
my $res = $node->safe_psql("postgres", qq(
SET enable_indexscan = off;
SELECT i FROM tst ORDER BY v <-> '$_' LIMIT $limit;
));
push(@expected, $res);
}
test_recall(0.18, $limit, "before vacuum");
test_recall(0.84, 100, "before vacuum");
# TODO Test concurrent inserts with vacuum
$node->safe_psql("postgres", "VACUUM tst;");
test_recall(0.84, $limit, "after vacuum");
done_testing();

View File

@@ -1,58 +0,0 @@
use strict;
use warnings FATAL => 'all';
use PostgreSQL::Test::Cluster;
use PostgreSQL::Test::Utils;
use Test::More;
# Initialize node
my $node = PostgreSQL::Test::Cluster->new('node');
$node->init;
$node->start;
# Create table
$node->safe_psql("postgres", "CREATE EXTENSION vector;");
$node->safe_psql("postgres", "CREATE TABLE tst (v intvec(3));");
sub insert_vectors
{
for my $i (1 .. 20)
{
$node->safe_psql("postgres", "INSERT INTO tst VALUES ('[1,1,1]');");
}
}
sub test_duplicates
{
my $res = $node->safe_psql("postgres", qq(
SET enable_seqscan = off;
SET hnsw.ef_search = 1;
SELECT COUNT(*) FROM (SELECT * FROM tst ORDER BY v <-> '[1,1,1]') t;
));
is($res, 10);
}
# Test duplicates with build
insert_vectors();
$node->safe_psql("postgres", "CREATE INDEX idx ON tst USING hnsw (v intvec_l2_ops);");
test_duplicates();
# Reset
$node->safe_psql("postgres", "TRUNCATE tst;");
# Test duplicates with inserts
insert_vectors();
test_duplicates();
# Test fallback path for inserts
$node->pgbench(
"--no-vacuum --client=5 --transactions=100",
0,
[qr{actually processed}],
[qr{^$}],
"concurrent INSERTs",
{
"042_hnsw_intvec_duplicates" => "INSERT INTO tst VALUES ('[1,1,1]');"
}
);
done_testing();