Compare commits

..

19 Commits

Author SHA1 Message Date
Andrew Kane
f6edf9245a Added support for adaptive beam search [skip ci] 2025-12-05 15:26:49 -08:00
Andrew Kane
b7c0085afe Added instructions for APK [skip ci] 2025-12-01 22:52:23 -08:00
Andrew Kane
2b742576b4 Added links [skip ci] 2025-11-24 12:41:30 -08:00
Andrew Kane
a126c02184 Added link to pgvector-ada [skip ci] 2025-10-30 22:29:36 -07:00
Andrew Kane
744362305b Updated changelog [skip ci] 2025-10-30 11:48:44 -07:00
Andrew Kane
69e78e36a3 Added link to pgvetor-pascal [skip ci] 2025-10-30 11:40:29 -07:00
Andrew Kane
8dd37d4fb2 Improved install target for Windows - #917 [skip ci] 2025-10-30 11:03:59 -07:00
Andrew Kane
a24125839a Updated readme [skip ci] 2025-10-28 10:25:34 -07:00
Andrew Kane
2ee3113417 Updated readme [skip ci] 2025-10-22 13:10:58 -07:00
Andrew Kane
3ebb9a506e Added varatt headers for Postgres 16+ 2025-10-22 11:56:05 -07:00
Andrew Kane
727d4836d2 Fixed Index Searches in EXPLAIN output for Postgres 18 2025-10-22 11:47:11 -07:00
Andrew Kane
d823c44591 Removed unused parameter [skip ci] 2025-09-27 15:58:18 -07:00
Andrew Kane
8f4aa0751c Removed unused parameters [skip ci] 2025-09-27 15:51:32 -07:00
Andrew Kane
1ff30720c7 Updated readme [skip ci] 2025-09-26 13:02:37 -07:00
Andrew Kane
4cef9213ec Added Docker images for Postgres 18 [skip ci] 2025-09-25 20:00:10 -07:00
Andrew Kane
90add68d6f Updated style to be consistent with Mac [skip ci] 2025-09-25 15:55:30 -07:00
Andrew Kane
bc05dbf312 Updated CI for Postgres 18 [skip ci] 2025-09-25 15:54:37 -07:00
Andrew Kane
02fefc0e3f Updated readme for Postgres 18 [skip ci] 2025-09-25 15:46:28 -07:00
Andrew Kane
e7899b1250 Updated checkout action [skip ci] 2025-09-22 11:55:01 -07:00
16 changed files with 111 additions and 93 deletions

View File

@@ -23,14 +23,14 @@ jobs:
- postgres: 13 - postgres: 13
os: ubuntu-22.04 os: ubuntu-22.04
steps: steps:
- uses: actions/checkout@v4 - uses: actions/checkout@v5
- uses: ankane/setup-postgres@v1 - uses: ankane/setup-postgres@v1
with: with:
postgres-version: ${{ matrix.postgres }} postgres-version: ${{ matrix.postgres }}
dev-files: true dev-files: true
- run: make - run: make
env: env:
PG_CFLAGS: ${{ matrix.postgres >= 18 && '-Wno-missing-field-initializers' || '' }} -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare ${{ matrix.postgres >= 18 && '-Wno-missing-field-initializers' || '' }}
- run: | - run: |
export PG_CONFIG=`which pg_config` export PG_CONFIG=`which pg_config`
sudo --preserve-env=PG_CONFIG make install sudo --preserve-env=PG_CONFIG make install
@@ -48,18 +48,18 @@ jobs:
fail-fast: false fail-fast: false
matrix: matrix:
include: include:
- postgres: 17 - postgres: 18
os: macos-26 os: macos-26
- postgres: 14 - postgres: 14
os: macos-15-intel os: macos-15-intel
steps: steps:
- uses: actions/checkout@v4 - uses: actions/checkout@v5
- uses: ankane/setup-postgres@v1 - uses: ankane/setup-postgres@v1
with: with:
postgres-version: ${{ matrix.postgres }} postgres-version: ${{ matrix.postgres }}
- run: make - run: make
env: env:
PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-unknown-warning-option PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-unknown-warning-option ${{ matrix.postgres >= 18 && '-Wno-missing-field-initializers' || '' }}
- run: make install - run: make install
- run: make installcheck - run: make installcheck
- if: ${{ failure() }} - if: ${{ failure() }}
@@ -72,7 +72,7 @@ jobs:
tar xf $TAG.tar.gz tar xf $TAG.tar.gz
mv postgres-$TAG postgres mv postgres-$TAG postgres
env: env:
TAG: ${{ matrix.postgres == 17 && 'REL_17_6' || 'REL_14_19' }} TAG: ${{ matrix.postgres == 18 && 'REL_18_0' || 'REL_14_19' }}
- run: make prove_installcheck PROVE_FLAGS="-I ./postgres/src/test/perl -I ./test/perl" - run: make prove_installcheck PROVE_FLAGS="-I ./postgres/src/test/perl -I ./test/perl"
env: env:
PERL5LIB: /Users/runner/perl5/lib/perl5 PERL5LIB: /Users/runner/perl5/lib/perl5
@@ -92,7 +92,7 @@ jobs:
- postgres: 14 - postgres: 14
os: windows-2022 os: windows-2022
steps: steps:
- uses: actions/checkout@v4 - uses: actions/checkout@v5
- uses: ankane/setup-postgres@v1 - uses: ankane/setup-postgres@v1
with: with:
postgres-version: ${{ matrix.postgres }} postgres-version: ${{ matrix.postgres }}
@@ -133,10 +133,10 @@ jobs:
if: ${{ !startsWith(github.ref_name, 'mac') && !startsWith(github.ref_name, 'windows') }} if: ${{ !startsWith(github.ref_name, 'mac') && !startsWith(github.ref_name, 'windows') }}
runs-on: ubuntu-latest runs-on: ubuntu-latest
steps: steps:
- uses: actions/checkout@v4 - uses: actions/checkout@v5
- uses: ankane/setup-postgres-valgrind@v1 - uses: ankane/setup-postgres-valgrind@v1
with: with:
postgres-version: 17 postgres-version: 18
check-ub: yes check-ub: yes
- run: make OPTFLAGS="" - run: make OPTFLAGS=""
- run: sudo --preserve-env=PG_CONFIG make install - run: sudo --preserve-env=PG_CONFIG make install

View File

@@ -1,3 +1,8 @@
## 0.8.2 (unreleased)
- Improved `install` target on Windows
- Fixed `Index Searches` in `EXPLAIN` output for Postgres 18
## 0.8.1 (2025-09-04) ## 0.8.1 (2025-09-04)
- Added support for Postgres 18 rc1 - Added support for Postgres 18 rc1

View File

@@ -57,7 +57,7 @@ install: all
copy $(SHLIB) "$(PKGLIBDIR)" copy $(SHLIB) "$(PKGLIBDIR)"
copy $(EXTENSION).control "$(SHAREDIR)\extension" copy $(EXTENSION).control "$(SHAREDIR)\extension"
copy sql\$(EXTENSION)--*.sql "$(SHAREDIR)\extension" copy sql\$(EXTENSION)--*.sql "$(SHAREDIR)\extension"
mkdir "$(INCLUDEDIR_SERVER)\extension\$(EXTENSION)" if not exist "$(INCLUDEDIR_SERVER)\extension\$(EXTENSION)" mkdir "$(INCLUDEDIR_SERVER)\extension\$(EXTENSION)"
for %f in ($(HEADERS)) do copy %f "$(INCLUDEDIR_SERVER)\extension\$(EXTENSION)" for %f in ($(HEADERS)) do copy %f "$(INCLUDEDIR_SERVER)\extension\$(EXTENSION)"
installcheck: installcheck:

View File

@@ -29,14 +29,14 @@ make install # may need sudo
See the [installation notes](#installation-notes---linux-and-mac) if you run into issues See the [installation notes](#installation-notes---linux-and-mac) if you run into issues
You can also install it with [Docker](#docker), [Homebrew](#homebrew), [PGXN](#pgxn), [APT](#apt), [Yum](#yum), [pkg](#pkg), or [conda-forge](#conda-forge), and it comes preinstalled with [Postgres.app](#postgresapp) and many [hosted providers](#hosted-postgres). There are also instructions for [GitHub Actions](https://github.com/pgvector/setup-pgvector). You can also install it with [Docker](#docker), [Homebrew](#homebrew), [PGXN](#pgxn), [APT](#apt), [Yum](#yum), [pkg](#pkg), [APK](#apk), or [conda-forge](#conda-forge), and it comes preinstalled with [Postgres.app](#postgresapp) and many [hosted providers](#hosted-postgres). There are also instructions for [GitHub Actions](https://github.com/pgvector/setup-pgvector).
### Windows ### Windows
Ensure [C++ support in Visual Studio](https://learn.microsoft.com/en-us/cpp/build/building-on-the-command-line?view=msvc-170#download-and-install-the-tools) is installed and run `x64 Native Tools Command Prompt for VS [version]` as administrator. Then use `nmake` to build: Ensure [C++ support in Visual Studio](https://learn.microsoft.com/en-us/cpp/build/building-on-the-command-line?view=msvc-170#download-and-install-the-tools) is installed and run `x64 Native Tools Command Prompt for VS [version]` as administrator. Then use `nmake` to build:
```cmd ```cmd
set "PGROOT=C:\Program Files\PostgreSQL\17" set "PGROOT=C:\Program Files\PostgreSQL\18"
cd %TEMP% cd %TEMP%
git clone --branch v0.8.1 https://github.com/pgvector/pgvector.git git clone --branch v0.8.1 https://github.com/pgvector/pgvector.git
cd pgvector cd pgvector
@@ -695,10 +695,10 @@ CREATE INDEX CONCURRENTLY ...
### Querying ### Querying
Use `EXPLAIN ANALYZE` to debug performance. Use `EXPLAIN (ANALYZE, BUFFERS)` to debug performance.
```sql ```sql
EXPLAIN ANALYZE SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5; EXPLAIN (ANALYZE, BUFFERS) SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
``` ```
#### Exact Search #### Exact Search
@@ -771,6 +771,8 @@ Use pgvector from any language with a Postgres client. You can even generate and
Language | Libraries / Examples Language | Libraries / Examples
--- | --- --- | ---
Ada | [pgvector-ada](https://github.com/pgvector/pgvector-ada)
Algol | [pgvector-algol](https://github.com/pgvector/pgvector-algol)
C | [pgvector-c](https://github.com/pgvector/pgvector-c) C | [pgvector-c](https://github.com/pgvector/pgvector-c)
C++ | [pgvector-cpp](https://github.com/pgvector/pgvector-cpp) C++ | [pgvector-cpp](https://github.com/pgvector/pgvector-cpp)
C#, F#, Visual Basic | [pgvector-dotnet](https://github.com/pgvector/pgvector-dotnet) C#, F#, Visual Basic | [pgvector-dotnet](https://github.com/pgvector/pgvector-dotnet)
@@ -790,10 +792,13 @@ Lisp | [pgvector-lisp](https://github.com/pgvector/pgvector-lisp)
Lua | [pgvector-lua](https://github.com/pgvector/pgvector-lua) Lua | [pgvector-lua](https://github.com/pgvector/pgvector-lua)
Nim | [pgvector-nim](https://github.com/pgvector/pgvector-nim) Nim | [pgvector-nim](https://github.com/pgvector/pgvector-nim)
OCaml | [pgvector-ocaml](https://github.com/pgvector/pgvector-ocaml) OCaml | [pgvector-ocaml](https://github.com/pgvector/pgvector-ocaml)
Pascal | [pgvector-pascal](https://github.com/pgvector/pgvector-pascal)
Perl | [pgvector-perl](https://github.com/pgvector/pgvector-perl) Perl | [pgvector-perl](https://github.com/pgvector/pgvector-perl)
PHP | [pgvector-php](https://github.com/pgvector/pgvector-php) PHP | [pgvector-php](https://github.com/pgvector/pgvector-php)
Prolog | [pgvector-prolog](https://github.com/pgvector/pgvector-prolog)
Python | [pgvector-python](https://github.com/pgvector/pgvector-python) Python | [pgvector-python](https://github.com/pgvector/pgvector-python)
R | [pgvector-r](https://github.com/pgvector/pgvector-r) R | [pgvector-r](https://github.com/pgvector/pgvector-r)
Racket | [pgvector-racket](https://github.com/pgvector/pgvector-racket)
Raku | [pgvector-raku](https://github.com/pgvector/pgvector-raku) Raku | [pgvector-raku](https://github.com/pgvector/pgvector-raku)
Ruby | [pgvector-ruby](https://github.com/pgvector/pgvector-ruby), [Neighbor](https://github.com/ankane/neighbor) Ruby | [pgvector-ruby](https://github.com/pgvector/pgvector-ruby), [Neighbor](https://github.com/ankane/neighbor)
Rust | [pgvector-rust](https://github.com/pgvector/pgvector-rust) Rust | [pgvector-rust](https://github.com/pgvector/pgvector-rust)
@@ -812,7 +817,7 @@ Yes, pgvector uses the write-ahead log (WAL), which allows for replication and p
#### What if I want to index vectors with more than 2,000 dimensions? #### What if I want to index vectors with more than 2,000 dimensions?
You can use [half-precision indexing](#half-precision-indexing) to index up to 4,000 dimensions or [binary quantization](#binary-quantization) to index up to 64,000 dimensions. Another option is [dimensionality reduction](https://en.wikipedia.org/wiki/Dimensionality_reduction). You can use [half-precision vectors](#half-precision-vectors) or [half-precision indexing](#half-precision-indexing) to index up to 4,000 dimensions or [binary quantization](#binary-quantization) to index up to 64,000 dimensions. Other options are [indexing subvectors](#indexing-subvectors) (for models that support it) or [dimensionality reduction](https://en.wikipedia.org/wiki/Dimensionality_reduction).
#### Can I store vectors with different dimensions in the same column? #### Can I store vectors with different dimensions in the same column?
@@ -1064,7 +1069,7 @@ 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: If your machine has multiple Postgres installations, specify the path to [pg_config](https://www.postgresql.org/docs/current/app-pgconfig.html) with:
```sh ```sh
export PG_CONFIG=/Library/PostgreSQL/17/bin/pg_config export PG_CONFIG=/Library/PostgreSQL/18/bin/pg_config
``` ```
Then re-run the installation instructions (run `make clean` before `make` if needed). If `sudo` is needed for `make install`, use: Then re-run the installation instructions (run `make clean` before `make` if needed). If `sudo` is needed for `make install`, use:
@@ -1075,11 +1080,11 @@ sudo --preserve-env=PG_CONFIG make install
A few common paths on Mac are: A few common paths on Mac are:
- EDB installer - `/Library/PostgreSQL/17/bin/pg_config` - EDB installer - `/Library/PostgreSQL/18/bin/pg_config`
- Homebrew (arm64) - `/opt/homebrew/opt/postgresql@17/bin/pg_config` - Homebrew (arm64) - `/opt/homebrew/opt/postgresql@18/bin/pg_config`
- Homebrew (x86-64) - `/usr/local/opt/postgresql@17/bin/pg_config` - Homebrew (x86-64) - `/usr/local/opt/postgresql@18/bin/pg_config`
Note: Replace `17` with your Postgres server version Note: Replace `18` with your Postgres server version
### Missing Header ### Missing Header
@@ -1088,10 +1093,10 @@ If compilation fails with `fatal error: postgres.h: No such file or directory`,
For Ubuntu and Debian, use: For Ubuntu and Debian, use:
```sh ```sh
sudo apt install postgresql-server-dev-17 sudo apt install postgresql-server-dev-18
``` ```
Note: Replace `17` with your Postgres server version Note: Replace `18` with your Postgres server version
### Missing SDK ### Missing SDK
@@ -1138,13 +1143,15 @@ If installation fails with `Access is denied`, re-run the installation instructi
Get the [Docker image](https://hub.docker.com/r/pgvector/pgvector) with: Get the [Docker image](https://hub.docker.com/r/pgvector/pgvector) with:
```sh ```sh
docker pull pgvector/pgvector:pg17-trixie docker pull pgvector/pgvector:pg18-trixie
``` ```
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 `18` with your Postgres server version, and run it the same way).
Supported tags are: Supported tags are:
- `pg18-trixie`, `0.8.1-pg18-trixie`
- `pg18-bookworm`, `0.8.1-pg18-bookworm`, `pg18`, `0.8.1-pg18`
- `pg17-trixie`, `0.8.1-pg17-trixie` - `pg17-trixie`, `0.8.1-pg17-trixie`
- `pg17-bookworm`, `0.8.1-pg17-bookworm`, `pg17`, `0.8.1-pg17` - `pg17-bookworm`, `0.8.1-pg17-bookworm`, `pg17`, `0.8.1-pg17`
- `pg16-trixie`, `0.8.1-pg16-trixie` - `pg16-trixie`, `0.8.1-pg16-trixie`
@@ -1161,7 +1168,7 @@ You can also build the image manually:
```sh ```sh
git clone --branch v0.8.1 https://github.com/pgvector/pgvector.git git clone --branch v0.8.1 https://github.com/pgvector/pgvector.git
cd pgvector cd pgvector
docker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector . docker build --pull --build-arg PG_MAJOR=18 -t myuser/pgvector .
``` ```
If you increase `maintenance_work_mem`, make sure `--shm-size` is at least that size to avoid an error with parallel HNSW index builds. If you increase `maintenance_work_mem`, make sure `--shm-size` is at least that size to avoid an error with parallel HNSW index builds.
@@ -1178,7 +1185,7 @@ With Homebrew Postgres, you can use:
brew install pgvector brew install pgvector
``` ```
Note: This only adds it to the `postgresql@17` and `postgresql@14` formulas Note: This only adds it to the `postgresql@18` and `postgresql@17` formulas
### PGXN ### PGXN
@@ -1193,22 +1200,22 @@ 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: 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 ```sh
sudo apt install postgresql-17-pgvector sudo apt install postgresql-18-pgvector
``` ```
Note: Replace `17` with your Postgres server version Note: Replace `18` with your Postgres server version
### Yum ### 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: 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 ```sh
sudo yum install pgvector_17 sudo yum install pgvector_18
# or # or
sudo dnf install pgvector_17 sudo dnf install pgvector_18
``` ```
Note: Replace `17` with your Postgres server version Note: Replace `18` with your Postgres server version
### pkg ### pkg
@@ -1225,6 +1232,14 @@ cd /usr/ports/databases/pgvector
make install make install
``` ```
### APK
Install the Alpine package with:
```sh
apk add postgresql-pgvector
```
### conda-forge ### conda-forge
With Conda Postgres, install from [conda-forge](https://anaconda.org/conda-forge/pgvector) with: With Conda Postgres, install from [conda-forge](https://anaconda.org/conda-forge/pgvector) with:
@@ -1267,6 +1282,7 @@ Thanks to:
- [k-means++: The Advantage of Careful Seeding](https://theory.stanford.edu/~sergei/papers/kMeansPP-soda.pdf) - [k-means++: The Advantage of Careful Seeding](https://theory.stanford.edu/~sergei/papers/kMeansPP-soda.pdf)
- [Concept Decompositions for Large Sparse Text Data using Clustering](https://www.cs.utexas.edu/users/inderjit/public_papers/concept_mlj.pdf) - [Concept Decompositions for Large Sparse Text Data using Clustering](https://www.cs.utexas.edu/users/inderjit/public_papers/concept_mlj.pdf)
- [Efficient and Robust Approximate Nearest Neighbor Search using Hierarchical Navigable Small World Graphs](https://arxiv.org/ftp/arxiv/papers/1603/1603.09320.pdf) - [Efficient and Robust Approximate Nearest Neighbor Search using Hierarchical Navigable Small World Graphs](https://arxiv.org/ftp/arxiv/papers/1603/1603.09320.pdf)
- [Distance Adaptive Beam Search for Provably Accurate Graph-Based Nearest Neighbor Search](https://arxiv.org/pdf/2505.15636)
## History ## History

View File

@@ -54,6 +54,10 @@
#include "utils/datum.h" #include "utils/datum.h"
#include "utils/memutils.h" #include "utils/memutils.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#if PG_VERSION_NUM >= 140000 #if PG_VERSION_NUM >= 140000
#include "utils/backend_progress.h" #include "utils/backend_progress.h"
#else #else
@@ -398,7 +402,7 @@ UpdateNeighborsInMemory(char *base, HnswSupport * support, HnswElement e, int m)
* Update graph in memory * Update graph in memory
*/ */
static void static void
UpdateGraphInMemory(HnswSupport * support, HnswElement element, int m, int efConstruction, HnswElement entryPoint, HnswBuildState * buildstate) UpdateGraphInMemory(HnswSupport * support, HnswElement element, int m, HnswElement entryPoint, HnswBuildState * buildstate)
{ {
HnswGraph *graph = buildstate->graph; HnswGraph *graph = buildstate->graph;
char *base = buildstate->hnswarea; char *base = buildstate->hnswarea;
@@ -460,7 +464,7 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
HnswFindElementNeighbors(base, element, entryPoint, NULL, support, m, efConstruction, false); HnswFindElementNeighbors(base, element, entryPoint, NULL, support, m, efConstruction, false);
/* Update graph in memory */ /* Update graph in memory */
UpdateGraphInMemory(support, element, m, efConstruction, entryPoint, buildstate); UpdateGraphInMemory(support, element, m, entryPoint, buildstate);
/* Release entry lock */ /* Release entry lock */
LWLockRelease(entryLock); LWLockRelease(entryLock);
@@ -1054,7 +1058,7 @@ ComputeParallelWorkers(Relation heap, Relation index)
* Build graph * Build graph
*/ */
static void static void
BuildGraph(HnswBuildState * buildstate, ForkNumber forkNum) BuildGraph(HnswBuildState * buildstate)
{ {
int parallel_workers = 0; int parallel_workers = 0;
@@ -1102,7 +1106,7 @@ BuildIndex(Relation heap, Relation index, IndexInfo *indexInfo,
InitBuildState(buildstate, heap, index, indexInfo, forkNum); InitBuildState(buildstate, heap, index, indexInfo, forkNum);
BuildGraph(buildstate, forkNum); BuildGraph(buildstate);
if (RelationNeedsWAL(index) || forkNum == INIT_FORKNUM) if (RelationNeedsWAL(index) || forkNum == INIT_FORKNUM)
log_newpage_range(index, forkNum, 0, RelationGetNumberOfBlocksInFork(index, forkNum), true); log_newpage_range(index, forkNum, 0, RelationGetNumberOfBlocksInFork(index, forkNum), true);

View File

@@ -9,6 +9,10 @@
#include "utils/datum.h" #include "utils/datum.h"
#include "utils/memutils.h" #include "utils/memutils.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
/* /*
* Get the insert page * Get the insert page
*/ */
@@ -660,7 +664,7 @@ FindDuplicateOnDisk(Relation index, HnswElement element, bool building)
* Update graph on disk * Update graph on disk
*/ */
static void static void
UpdateGraphOnDisk(Relation index, HnswSupport * support, HnswElement element, int m, int efConstruction, HnswElement entryPoint, bool building) UpdateGraphOnDisk(Relation index, HnswSupport * support, HnswElement element, int m, HnswElement entryPoint, bool building)
{ {
BlockNumber newInsertPage = InvalidBlockNumber; BlockNumber newInsertPage = InvalidBlockNumber;
@@ -728,7 +732,7 @@ HnswInsertTupleOnDisk(Relation index, HnswSupport * support, Datum value, ItemPo
HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, false); HnswFindElementNeighbors(base, element, entryPoint, index, support, m, efConstruction, false);
/* Update graph on disk */ /* Update graph on disk */
UpdateGraphOnDisk(index, support, element, m, efConstruction, entryPoint, building); UpdateGraphOnDisk(index, support, element, m, entryPoint, building);
/* Release lock */ /* Release lock */
UnlockPage(index, HNSW_UPDATE_LOCK, lockmode); UnlockPage(index, HNSW_UPDATE_LOCK, lockmode);

View File

@@ -193,6 +193,10 @@ hnswgettuple(IndexScanDesc scan, ScanDirection dir)
/* Count index scan for stats */ /* Count index scan for stats */
pgstat_count_index_scan(scan->indexRelation); pgstat_count_index_scan(scan->indexRelation);
#if PG_VERSION_NUM >= 180000
if (scan->instrument)
scan->instrument->nsearches++;
#endif
/* Safety check */ /* Safety check */
if (scan->orderByData == NULL) if (scan->orderByData == NULL)

View File

@@ -15,6 +15,10 @@
#include "utils/memdebug.h" #include "utils/memdebug.h"
#include "utils/rel.h" #include "utils/rel.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#if PG_VERSION_NUM < 170000 #if PG_VERSION_NUM < 170000
static inline uint64 static inline uint64
murmurhash64(uint64 data) murmurhash64(uint64 data)
@@ -825,6 +829,7 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
HnswUnvisited *unvisited = palloc(lm * sizeof(HnswUnvisited)); HnswUnvisited *unvisited = palloc(lm * sizeof(HnswUnvisited));
int unvisitedLength; int unvisitedLength;
bool inMemory = index == NULL; bool inMemory = index == NULL;
double lambda = 0.12;
if (v == NULL) if (v == NULL)
{ {
@@ -880,7 +885,7 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
HnswSearchCandidate *f = HnswGetSearchCandidate(w_node, pairingheap_first(W)); HnswSearchCandidate *f = HnswGetSearchCandidate(w_node, pairingheap_first(W));
HnswElement cElement; HnswElement cElement;
if (c->distance > f->distance) if (c->distance > f->distance * (1 + lambda))
break; break;
cElement = HnswPtrAccess(base, c->element); cElement = HnswPtrAccess(base, c->element);
@@ -900,8 +905,10 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
HnswSearchCandidate *e; HnswSearchCandidate *e;
double eDistance; double eDistance;
bool alwaysAdd = wlen < ef; bool alwaysAdd = wlen < ef;
double fDistance;
f = HnswGetSearchCandidate(w_node, pairingheap_first(W)); f = HnswGetSearchCandidate(w_node, pairingheap_first(W));
fDistance = f->distance * (1 + lambda);
if (inMemory) if (inMemory)
{ {
@@ -916,13 +923,13 @@ HnswSearchLayer(char *base, HnswQuery * q, List *ep, int ef, int lc, Relation in
/* Avoid any allocations if not adding */ /* Avoid any allocations if not adding */
eElement = NULL; eElement = NULL;
HnswLoadElementImpl(blkno, offno, &eDistance, q, index, support, inserting, alwaysAdd || discarded != NULL ? NULL : &f->distance, &eElement); HnswLoadElementImpl(blkno, offno, &eDistance, q, index, support, inserting, alwaysAdd || discarded != NULL ? NULL : &fDistance, &eElement);
if (eElement == NULL) if (eElement == NULL)
continue; continue;
} }
if (eElement == NULL || !(eDistance < f->distance || alwaysAdd)) if (eElement == NULL || !(eDistance < fDistance || alwaysAdd))
{ {
if (discarded != NULL) if (discarded != NULL)
{ {

View File

@@ -9,6 +9,10 @@
#include "storage/lmgr.h" #include "storage/lmgr.h"
#include "utils/memutils.h" #include "utils/memutils.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#if PG_VERSION_NUM >= 180000 #if PG_VERSION_NUM >= 180000
#define vacuum_delay_point() vacuum_delay_point(false) #define vacuum_delay_point() vacuum_delay_point(false)
#endif #endif

View File

@@ -20,6 +20,10 @@
#include "utils/memutils.h" #include "utils/memutils.h"
#include "vector.h" #include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#if PG_VERSION_NUM >= 140000 #if PG_VERSION_NUM >= 140000
#include "utils/backend_progress.h" #include "utils/backend_progress.h"
#else #else
@@ -138,7 +142,7 @@ SampleRows(IvfflatBuildState * buildstate)
* Add tuple to sort * Add tuple to sort
*/ */
static void static void
AddTupleToSort(Relation index, ItemPointer tid, Datum *values, IvfflatBuildState * buildstate) AddTupleToSort(ItemPointer tid, Datum *values, IvfflatBuildState * buildstate)
{ {
double distance; double distance;
double minDistance = DBL_MAX; double minDistance = DBL_MAX;
@@ -215,7 +219,7 @@ BuildCallback(Relation index, ItemPointer tid, Datum *values,
oldCtx = MemoryContextSwitchTo(buildstate->tmpCtx); oldCtx = MemoryContextSwitchTo(buildstate->tmpCtx);
/* Add tuple to sort */ /* Add tuple to sort */
AddTupleToSort(index, tid, values, buildstate); AddTupleToSort(tid, values, buildstate);
/* Reset memory context */ /* Reset memory context */
MemoryContextSwitchTo(oldCtx); MemoryContextSwitchTo(oldCtx);
@@ -470,8 +474,8 @@ CreateMetaPage(Relation index, int dimensions, int lists, ForkNumber forkNum)
* Create list pages * Create list pages
*/ */
static void static void
CreateListPages(Relation index, VectorArray centers, int dimensions, CreateListPages(Relation index, VectorArray centers, int lists,
int lists, ForkNumber forkNum, ListInfo * *listInfo) ForkNumber forkNum, ListInfo * *listInfo)
{ {
Buffer buf; Buffer buf;
Page page; Page page;
@@ -1004,7 +1008,7 @@ BuildIndex(Relation heap, Relation index, IndexInfo *indexInfo,
/* Create pages */ /* Create pages */
CreateMetaPage(index, buildstate->dimensions, buildstate->lists, forkNum); CreateMetaPage(index, buildstate->dimensions, buildstate->lists, forkNum);
CreateListPages(index, buildstate->centers, buildstate->dimensions, buildstate->lists, forkNum, &buildstate->listInfo); CreateListPages(index, buildstate->centers, buildstate->lists, forkNum, &buildstate->listInfo);
CreateEntryPages(buildstate, forkNum); CreateEntryPages(buildstate, forkNum);
/* Write WAL for initialization fork since GenericXLog functions do not */ /* Write WAL for initialization fork since GenericXLog functions do not */

View File

@@ -143,10 +143,6 @@ ivfflatcostestimate(PlannerInfo *root, IndexPath *path, double loop_count,
*indexSelectivity = costs.indexSelectivity; *indexSelectivity = costs.indexSelectivity;
*indexCorrelation = costs.indexCorrelation; *indexCorrelation = costs.indexCorrelation;
*indexPages = costs.numIndexPages; *indexPages = costs.numIndexPages;
elog(INFO, "ivfflatcostestimate = %f", costs.indexTotalCost);
/* Cost estimates for parallel workers applied outside of amcostestimate */
elog(INFO, "parallel_workers = %d, parallel aware = %d", path->path.parallel_workers, path->path.parallel_aware);
} }
/* /*
@@ -174,25 +170,6 @@ ivfflatvalidate(Oid opclassoid)
return true; return true;
} }
static Size
ivfflatestimateparallelscan()
{
elog(INFO, "ivfflatestimateparallelscan");
return 0;
}
static void
ivfflatinitparallelscan(void *target)
{
elog(INFO, "ivfflatinitparallelscan");
}
static void
ivfflatparallelrescan(IndexScanDesc scan)
{
elog(INFO, "ivfflatparallelrescan");
}
/* /*
* Define index handler * Define index handler
* *
@@ -223,7 +200,7 @@ ivfflathandler(PG_FUNCTION_ARGS)
amroutine->amstorage = false; amroutine->amstorage = false;
amroutine->amclusterable = false; amroutine->amclusterable = false;
amroutine->ampredlocks = false; amroutine->ampredlocks = false;
amroutine->amcanparallel = true; amroutine->amcanparallel = false;
#if PG_VERSION_NUM >= 170000 #if PG_VERSION_NUM >= 170000
amroutine->amcanbuildparallel = true; amroutine->amcanbuildparallel = true;
#endif #endif
@@ -265,9 +242,9 @@ ivfflathandler(PG_FUNCTION_ARGS)
amroutine->amrestrpos = NULL; amroutine->amrestrpos = NULL;
/* Interface functions to support parallel index scans */ /* Interface functions to support parallel index scans */
amroutine->amestimateparallelscan = ivfflatestimateparallelscan; amroutine->amestimateparallelscan = NULL;
amroutine->aminitparallelscan = ivfflatinitparallelscan; amroutine->aminitparallelscan = NULL;
amroutine->amparallelrescan = ivfflatparallelrescan; amroutine->amparallelrescan = NULL;
#if PG_VERSION_NUM >= 180000 #if PG_VERSION_NUM >= 180000
amroutine->amtranslatestrategy = NULL; amroutine->amtranslatestrategy = NULL;

View File

@@ -13,6 +13,10 @@
#include "utils/tuplesort.h" #include "utils/tuplesort.h"
#include "vector.h" #include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
#if PG_VERSION_NUM >= 150000 #if PG_VERSION_NUM >= 150000
#include "common/pg_prng.h" #include "common/pg_prng.h"
#endif #endif

View File

@@ -65,7 +65,7 @@ FindInsertPage(Relation index, Datum *values, BlockNumber *insertPage, ListInfo
* Insert a tuple into the index * Insert a tuple into the index
*/ */
static void static void
InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heap_tid, Relation heapRel) InsertTuple(Relation index, Datum *values, bool *isnull, ItemPointer heap_tid)
{ {
const IvfflatTypeInfo *typeInfo = IvfflatGetTypeInfo(index); const IvfflatTypeInfo *typeInfo = IvfflatGetTypeInfo(index);
IndexTuple itup; IndexTuple itup;
@@ -204,7 +204,7 @@ ivfflatinsert(Relation index, Datum *values, bool *isnull, ItemPointer heap_tid,
oldCtx = MemoryContextSwitchTo(insertCtx); oldCtx = MemoryContextSwitchTo(insertCtx);
/* Insert tuple */ /* Insert tuple */
InsertTuple(index, values, isnull, heap_tid, heap); InsertTuple(index, values, isnull, heap_tid);
/* Delete memory context */ /* Delete memory context */
MemoryContextSwitchTo(oldCtx); MemoryContextSwitchTo(oldCtx);

View File

@@ -13,6 +13,10 @@
#include "utils/memutils.h" #include "utils/memutils.h"
#include "vector.h" #include "vector.h"
#if PG_VERSION_NUM >= 160000
#include "varatt.h"
#endif
/* /*
* Initialize with kmeans++ * Initialize with kmeans++
* *

View File

@@ -119,9 +119,6 @@ GetScanItems(IndexScanDesc scan, Datum value)
tuplesort_reset(so->sortstate); tuplesort_reset(so->sortstate);
if (scan->parallel_scan != NULL)
elog(INFO, "parallel scan");
/* Search closest probes lists */ /* Search closest probes lists */
while (so->listIndex < so->maxProbes && (++batchProbes) <= so->probes) while (so->listIndex < so->maxProbes && (++batchProbes) <= so->probes)
{ {
@@ -358,6 +355,10 @@ ivfflatgettuple(IndexScanDesc scan, ScanDirection dir)
/* Count index scan for stats */ /* Count index scan for stats */
pgstat_count_index_scan(scan->indexRelation); pgstat_count_index_scan(scan->indexRelation);
#if PG_VERSION_NUM >= 180000
if (scan->instrument)
scan->instrument->nsearches++;
#endif
/* Safety check */ /* Safety check */
if (scan->orderByData == NULL) if (scan->orderByData == NULL)

View File

@@ -1,16 +0,0 @@
-- SET force_parallel_mode = on;
SET parallel_setup_cost = 10;
SET parallel_tuple_cost = 0.000001;
SET min_parallel_table_scan_size = 1;
SET min_parallel_index_scan_size = 1;
CREATE TABLE t (id integer, val vector(3));
ALTER TABLE t ALTER COLUMN val SET STORAGE PLAIN;
INSERT INTO t (id, val) SELECT n, ARRAY[random(), random(), random()] FROM generate_series(1,1000000) n;
CREATE INDEX ON t USING ivfflat (val) WITH (lists = 10);
SET ivfflat.probes = 4;
EXPLAIN SELECT * FROM t ORDER BY val <-> '[0.5,0.5,0.5]' LIMIT 5;
SELECT * FROM t ORDER BY val <-> '[0.5,0.5,0.5]' LIMIT 5;
DROP TABLE t;