Added subscript function for vectors

Redefining a typedef is a C11 feature:

    In file included from src/hnsw.c:10:
    src/hnsw.h:147:5: warning: redefinition of typedef 'HnswElementData' is a C11 feature [-Wtypedef-redefinition]
    }                       HnswElementData;
                            ^
    src/hnsw.h:118:32: note: previous definition is here
    typedef struct HnswElementData HnswElementData;
                                   ^
    src/hnsw.h:163:5: warning: redefinition of typedef 'HnswNeighborArray' is a C11 feature [-Wtypedef-redefinition]
    }                       HnswNeighborArray;
                            ^
    src/hnsw.h:119:34: note: previous definition is here
    typedef struct HnswNeighborArray HnswNeighborArray;
                                     ^
    2 warnings generated.

I got these warnings when I built PostgreSQL with "CC=clang
CFLAGS=-std=gnu99"; other similar options would surely produce the
warnings too.

.github/workflows/build.yml

@@ -28,7 +28,7 @@ jobs:
           dev-files: true
       - run: make
         env:
-          PG_CFLAGS: -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare
+          PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare
       - run: |
           export PG_CONFIG=`which pg_config`
           sudo --preserve-env=PG_CONFIG make install
@@ -49,7 +49,7 @@ jobs:
           postgres-version: 14
       - run: make
         env:
-          PG_CFLAGS: -Wall -Wextra -Werror -Wno-unused-parameter
+          PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter
       - run: make install
       - run: make installcheck
       - if: ${{ failure() }}
@@ -60,7 +60,9 @@ jobs:
           wget -q https://github.com/postgres/postgres/archive/refs/tags/REL_14_10.tar.gz
           tar xf REL_14_10.tar.gz
       - run: make prove_installcheck PROVE_FLAGS="-I ./postgres-REL_14_10/src/test/perl" PERL5LIB="/Users/runner/perl5/lib/perl5"
-      - run: make clean && /usr/local/opt/llvm@15/bin/scan-build --status-bugs make PG_CFLAGS="-DUSE_ASSERT_CHECKING"
+      - run: make clean && /usr/local/opt/llvm@15/bin/scan-build --status-bugs make
+        env:
+          PG_CFLAGS: -DUSE_ASSERT_CHECKING
   windows:
     runs-on: windows-latest
     if: ${{ !startsWith(github.ref_name, 'mac') }}
@@ -71,6 +73,7 @@ jobs:
           postgres-version: 14
       - run: |
           call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvars64.bat" && ^
+          cd %TEMP% && ^
           nmake /NOLOGO /F Makefile.win && ^
           nmake /NOLOGO /F Makefile.win install && ^
           nmake /NOLOGO /F Makefile.win installcheck && ^
@@ -97,4 +100,15 @@ jobs:
           sudo -u postgres make installcheck
           sudo -u postgres make prove_installcheck
         env:
-          PG_CFLAGS: -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare
+          PG_CFLAGS: -DUSE_ASSERT_CHECKING -Wall -Wextra -Werror -Wno-unused-parameter -Wno-sign-compare
+  valgrind:
+    if: ${{ !startsWith(github.ref_name, 'mac') && !startsWith(github.ref_name, 'windows') }}
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: ankane/setup-postgres-valgrind@v1
+        with:
+          postgres-version: 16
+      - run: make
+      - run: sudo --preserve-env=PG_CONFIG make install
+      - run: make installcheck

CHANGELOG.md

@@ -1,11 +1,25 @@
-## 0.6.0 (unreleased)
+## 0.7.0 (unreleased)
 
-If upgrading with Postgres 12 or Docker, see [these notes](https://github.com/pgvector/pgvector#060-unreleased).
+- Added subscript function for vectors
+
+## 0.6.2 (unreleased)
+
+- Reduced lock contention with parallel HNSW index builds
+
+## 0.6.1 (2024-03-04)
+
+- Fixed error with `ANALYZE` and vectors with different dimensions
+- Fixed error with `shared_preload_libraries`
+- Fixed vector subtraction being marked as commutative
+
+## 0.6.0 (2024-01-29)
+
+If upgrading with Postgres 12 or Docker, see [these notes](https://github.com/pgvector/pgvector#060).
 
-- Changed storage for vector from `extended` to `external`
-- Improved performance of HNSW
 - Added support for parallel index builds for HNSW
 - Added validation for GUC parameters
+- Changed storage for vector from `extended` to `external`
+- Improved performance of HNSW
 - Reduced memory usage for HNSW index builds
 - Reduced WAL generation for HNSW index builds
 - Fixed error with logical replication

Dockerfile

@@ -1,4 +1,4 @@
-ARG PG_MAJOR=15
+ARG PG_MAJOR=16
 FROM postgres:$PG_MAJOR
 ARG PG_MAJOR
 

LICENSE

@@ -1,4 +1,4 @@
-Portions Copyright (c) 1996-2023, PostgreSQL Global Development Group
+Portions Copyright (c) 1996-2024, PostgreSQL Global Development Group
 
 Portions Copyright (c) 1994, The Regents of the University of California
 

META.json

@@ -2,7 +2,7 @@
 	"name": "vector",
 	"abstract": "Open-source vector similarity search for Postgres",
 	"description": "Supports L2 distance, inner product, and cosine distance",
-	"version": "0.5.1",
+	"version": "0.6.1",
 	"maintainer": [
 		"Andrew Kane <andrew@ankane.org>"
 	],
@@ -20,7 +20,7 @@
 		"vector": {
 			"file": "sql/vector.sql",
 			"docfile": "README.md",
-			"version": "0.5.1",
+			"version": "0.6.1",
 			"abstract": "Open-source vector similarity search for Postgres"
 		}
 	},

Makefile

@@ -1,5 +1,5 @@
 EXTENSION = vector
-EXTVERSION = 0.5.1
+EXTVERSION = 0.6.1
 
 MODULE_big = vector
 DATA = $(wildcard sql/*--*.sql)
@@ -71,11 +71,9 @@ PG_MAJOR ?= 16
 .PHONY: docker
 
 docker:
-	docker build --pull --no-cache --build-arg PG_MAJOR=$(PG_MAJOR) -t pgvector/pgvector:pg$(PG_MAJOR) .
+	docker build --pull --no-cache --build-arg PG_MAJOR=$(PG_MAJOR) -t pgvector/pgvector:pg$(PG_MAJOR) -t pgvector/pgvector:$(EXTVERSION)-pg$(PG_MAJOR) .
-	docker build --build-arg PG_MAJOR=$(PG_MAJOR) -t pgvector/pgvector:$(EXTVERSION)-pg$(PG_MAJOR) .
 
 .PHONY: docker-release
 
 docker-release:
-	docker buildx build --push --pull --no-cache --platform linux/amd64,linux/arm64 --build-arg PG_MAJOR=$(PG_MAJOR) -t pgvector/pgvector:pg$(PG_MAJOR) .
+	docker buildx build --push --pull --no-cache --platform linux/amd64,linux/arm64 --build-arg PG_MAJOR=$(PG_MAJOR) -t pgvector/pgvector:pg$(PG_MAJOR) -t pgvector/pgvector:$(EXTVERSION)-pg$(PG_MAJOR) .
-	docker buildx build --push --platform linux/amd64,linux/arm64 --build-arg PG_MAJOR=$(PG_MAJOR) -t pgvector/pgvector:$(EXTVERSION)-pg$(PG_MAJOR) .

Makefile.win

@@ -1,5 +1,5 @@
 EXTENSION = vector
-EXTVERSION = 0.5.1
+EXTVERSION = 0.6.1
 
 OBJS = 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\vector.obj
 HEADERS = src\vector.h

README.md

@@ -16,19 +16,19 @@ Plus [ACID](https://en.wikipedia.org/wiki/ACID) compliance, point-in-time recove
 
 ### Linux and Mac
 
-Compile and install the extension (supports Postgres 11+)
+Compile and install the extension (supports Postgres 12+)
 
 ```sh
 cd /tmp
-git clone --branch v0.5.1 https://github.com/pgvector/pgvector.git
+git clone --branch v0.6.1 https://github.com/pgvector/pgvector.git
 cd pgvector
 make
 make install # may need sudo
 ```
 
-See the [installation notes](#installation-notes) 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), 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), 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
 
@@ -44,12 +44,15 @@ Then use `nmake` to build:
 
 ```cmd
 set "PGROOT=C:\Program Files\PostgreSQL\16"
-git clone --branch v0.5.1 https://github.com/pgvector/pgvector.git
+cd %TEMP%
+git clone --branch v0.6.1 https://github.com/pgvector/pgvector.git
 cd pgvector
 nmake /F Makefile.win
 nmake /F Makefile.win install
 ```
 
+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).
 
 ## Getting Started
@@ -102,6 +105,12 @@ Insert vectors
 INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]');
 ```
 
+Or load vectors in bulk using `COPY` ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/bulk_loading.py))
+
+```sql
+COPY items (embedding) FROM STDIN WITH (FORMAT BINARY);
+```
+
 Upsert vectors
 
 ```sql
@@ -264,7 +273,9 @@ HINT:  Increase maintenance_work_mem to speed up builds.
 
 Note: Do not set `maintenance_work_mem` so high that it exhausts the memory on the server
 
-Starting with 0.6.0 [unreleased], you can also speed up index creation by increasing the number of parallel workers (2 by default)
+Like other index types, it’s faster to create an index after loading your initial data
+
+Starting with 0.6.0, you can also speed up index creation by increasing the number of parallel workers (2 by default)
 
 ```sql
 SET max_parallel_maintenance_workers = 7; -- plus leader
@@ -402,13 +413,39 @@ You can use [Reciprocal Rank Fusion](https://github.com/pgvector/pgvector-python
 
 ## Performance
 
+### Tuning
+
+Use a tool like [PgTune](https://pgtune.leopard.in.ua/) to set initial values for Postgres server parameters.
+
+### Loading
+
+Use `COPY` for bulk loading data ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/bulk_loading.py)).
+
+```sql
+COPY items (embedding) FROM STDIN WITH (FORMAT BINARY);
+```
+
+Add any indexes *after* loading the initial data for best performance.
+
+### Indexing
+
+See index build time for [HNSW](#index-build-time) and [IVFFlat](#index-build-time-1).
+
+In production environments, create indexes concurrently to avoid blocking writes.
+
+```sql
+CREATE INDEX CONCURRENTLY ...
+```
+
+### Querying
+
 Use `EXPLAIN ANALYZE` to debug performance.
 
 ```sql
 EXPLAIN ANALYZE SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;
 ```
 
-### Exact Search
+#### Exact Search
 
 To speed up queries without an index, increase `max_parallel_workers_per_gather`.
 
@@ -422,7 +459,7 @@ If vectors are normalized to length 1 (like [OpenAI embeddings](https://platform
 SELECT * FROM items ORDER BY embedding <#> '[3,1,2]' LIMIT 5;
 ```
 
-### Approximate Search
+#### Approximate Search
 
 To speed up queries with an IVFFlat index, increase the number of inverted lists (at the expense of recall).
 
@@ -430,6 +467,50 @@ To speed up queries with an IVFFlat index, increase the number of inverted lists
 CREATE INDEX ON items USING ivfflat (embedding vector_l2_ops) WITH (lists = 1000);
 ```
 
+### Vacuuming
+
+Vacuuming can take a while for HNSW indexes. Speed it up by reindexing first.
+
+```sql
+REINDEX INDEX CONCURRENTLY index_name;
+VACUUM table_name;
+```
+
+## Monitoring
+
+Monitor performance with [pg_stat_statements](https://www.postgresql.org/docs/current/pgstatstatements.html) (be sure to add it to `shared_preload_libraries`).
+
+```sql
+CREATE EXTENSION pg_stat_statements;
+```
+
+Get the most time-consuming queries with:
+
+```sql
+SELECT query, calls, ROUND((total_plan_time + total_exec_time) / calls) AS avg_time_ms,
+    ROUND((total_plan_time + total_exec_time) / 60000) AS total_time_min
+    FROM pg_stat_statements ORDER BY total_plan_time + total_exec_time DESC LIMIT 20;
+```
+
+Note: Replace `total_plan_time + total_exec_time` with `total_time` for Postgres < 13
+
+Monitor recall by comparing results from approximate search with exact search.
+
+```sql
+BEGIN;
+SET LOCAL enable_indexscan = off; -- use exact search
+SELECT ...
+COMMIT;
+```
+
+## Scaling
+
+Scale pgvector the same way you scale Postgres.
+
+Scale vertically by increasing memory, CPU, and storage on a single instance. Use existing tools to [tune parameters](#tuning) and [monitor performance](#monitoring).
+
+Scale horizontally with [replicas](https://www.postgresql.org/docs/current/hot-standby.html), or use [Citus](https://github.com/citusdata/citus) or another approach for sharding ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/citus.py)).
+
 ## Languages
 
 Use pgvector from any language with a Postgres client. You can even generate and store vectors in one language and query them in another.
@@ -535,7 +616,17 @@ SELECT pg_size_pretty(pg_relation_size('index_name'));
 
 #### Why isn’t a query using an index?
 
-The cost estimation in pgvector < 0.4.3 does not always work well with the planner. You can encourage the planner to use an index for a query with:
+The query needs to have an `ORDER BY` and `LIMIT`, and the `ORDER BY` must be the result of a distance operator, not an expression.
+
+```sql
+-- index
+ORDER BY embedding <=> '[3,1,2]' LIMIT 5;
+
+-- no index
+ORDER BY 1 - (embedding <=> '[3,1,2]') DESC LIMIT 5;
+```
+
+You can encourage the planner to use an index for a query with:
 
 ```sql
 BEGIN;
@@ -564,6 +655,10 @@ or choose to store vectors inline:
 ALTER TABLE items ALTER COLUMN embedding SET STORAGE PLAIN;
 ```
 
+#### Why are there less results for a query after adding an HNSW index?
+
+Results are limited by the size of the dynamic candidate list (`hnsw.ef_search`). There may be even less results due to dead tuples or filtering conditions in the query. We recommend setting `hnsw.ef_search` to at least twice the `LIMIT` of the query. If you need more than 500 results, use an IVFFlat index instead.
+
 #### Why are there less results for a query after adding an IVFFlat index?
 
 The index was likely created with too little data for the number of lists. Drop the index until the table has more data.
@@ -572,6 +667,8 @@ The index was likely created with too little data for the number of lists. Drop
 DROP INDEX index_name;
 ```
 
+Results can also be limited by the number of probes (`ivfflat.probes`).
+
 ## Reference
 
 ### Vector Type
@@ -607,7 +704,7 @@ Function | Description | Added
 avg(vector) → vector | average |
 sum(vector) → vector | sum | 0.5.0
 
-## Installation Notes
+## Installation Notes - Linux and Mac
 
 ### Postgres Location
 
@@ -647,24 +744,44 @@ Note: Replace `16` with your Postgres server version
 
 If compilation fails and the output includes `warning: no such sysroot directory` on Mac, reinstall Xcode Command Line Tools.
 
+### Portability
+
+By default, pgvector compiles with `-march=native` on some platforms for best performance. However, this can lead to `Illegal instruction` errors if trying to run the compiled extension on a different machine.
+
+To compile for portability, use:
+
+```sh
+make OPTFLAGS=""
+```
+
+## Installation Notes - Windows
+
+### Missing Header
+
+If compilation fails with `Cannot open include file: 'postgres.h': No such file or directory`, make sure `PGROOT` is correct.
+
+### Permissions
+
+If installation fails with `Access is denied`, re-run the installation instructions as an administrator.
+
 ## Additional Installation Methods
 
 ### Docker
 
-Get the [Docker image](https://hub.docker.com/r/ankane/pgvector) with:
+Get the [Docker image](https://hub.docker.com/r/pgvector/pgvector) with:
 
 ```sh
-docker pull ankane/pgvector
+docker pull pgvector/pgvector:pg16
 ```
 
-This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (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 v0.5.1 https://github.com/pgvector/pgvector.git
+git clone --branch v0.6.1 https://github.com/pgvector/pgvector.git
 cd pgvector
-docker build --build-arg PG_MAJOR=15 -t myuser/pgvector .
+docker build --build-arg PG_MAJOR=16 -t myuser/pgvector .
 ```
 
 ### Homebrew
@@ -707,6 +824,21 @@ sudo dnf install pgvector_16
 
 Note: Replace `16` with your Postgres server version
 
+### pkg
+
+Install the FreeBSD package with:
+
+```sh
+pkg install postgresql15-pg_vector
+```
+
+or the port with:
+
+```sh
+cd /usr/ports/databases/pgvector
+make install
+```
+
 ### conda-forge
 
 With Conda Postgres, install from [conda-forge](https://anaconda.org/conda-forge/pgvector) with:
@@ -741,7 +873,7 @@ SELECT extversion FROM pg_extension WHERE extname = 'vector';
 
 ## Upgrade Notes
 
-### 0.6.0 [unreleased]
+### 0.6.0
 
 #### Postgres 12
 
@@ -816,6 +948,12 @@ make installcheck REGRESS=functions                            # regression test
 make prove_installcheck PROVE_TESTS=test/t/001_ivfflat_wal.pl  # TAP test
 ```
 
+To enable assertions:
+
+```sh
+make clean && PG_CFLAGS="-DUSE_ASSERT_CHECKING" make && make install
+```
+
 To enable benchmarking:
 
 ```sh
@@ -828,12 +966,6 @@ To show memory usage:
 make clean && PG_CFLAGS="-DHNSW_MEMORY -DIVFFLAT_MEMORY" make && make install
 ```
 
-To enable assertions:
-
-```sh
-make clean && PG_CFLAGS="-DUSE_ASSERT_CHECKING" make && make install
-```
-
 To get k-means metrics:
 
 ```sh

sql/vector--0.6.0--0.6.1.sql

@@ -0,0 +1,16 @@
+-- complain if script is sourced in psql, rather than via CREATE EXTENSION
+\echo Use "ALTER EXTENSION vector UPDATE TO '0.6.1'" to load this file. \quit
+
+DROP OPERATOR - (vector, vector);
+
+CREATE OPERATOR - (
+	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_sub
+);
+
+ALTER OPERATOR <= (vector, vector) SET (
+	RESTRICT = scalarlesel, JOIN = scalarlejoinsel
+);
+
+ALTER OPERATOR >= (vector, vector) SET (
+	RESTRICT = scalargesel, JOIN = scalargejoinsel
+);

sql/vector--0.6.2--0.7.0.sql

@@ -0,0 +1,7 @@
+-- complain if script is sourced in psql, rather than via CREATE EXTENSION
+\echo Use "ALTER EXTENSION vector UPDATE TO '0.7.0'" to load this file. \quit
+
+CREATE FUNCTION vector_subscript(internal) RETURNS internal
+	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
+
+ALTER TYPE vector SET (SUBSCRIPT = vector_subscript);

sql/vector.sql

@@ -20,12 +20,16 @@ CREATE FUNCTION vector_recv(internal, oid, integer) RETURNS vector
 CREATE FUNCTION vector_send(vector) RETURNS bytea
 	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
 
+CREATE FUNCTION vector_subscript(internal) RETURNS internal
+	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
+
 CREATE TYPE vector (
 	INPUT     = vector_in,
 	OUTPUT    = vector_out,
 	TYPMOD_IN = vector_typmod_in,
 	RECEIVE   = vector_recv,
 	SEND      = vector_send,
+	SUBSCRIPT = vector_subscript,
 	STORAGE   = external
 );
 
@@ -180,8 +184,7 @@ CREATE OPERATOR + (
 );
 
 CREATE OPERATOR - (
-	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_sub,
+	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_sub
-	COMMUTATOR = -
 );
 
 CREATE OPERATOR * (
@@ -195,11 +198,10 @@ CREATE OPERATOR < (
 	RESTRICT = scalarltsel, JOIN = scalarltjoinsel
 );
 
--- should use scalarlesel and scalarlejoinsel, but not supported in Postgres < 11
 CREATE OPERATOR <= (
 	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_le,
 	COMMUTATOR = >= , NEGATOR = > ,
-	RESTRICT = scalarltsel, JOIN = scalarltjoinsel
+	RESTRICT = scalarlesel, JOIN = scalarlejoinsel
 );
 
 CREATE OPERATOR = (
@@ -214,11 +216,10 @@ CREATE OPERATOR <> (
 	RESTRICT = eqsel, JOIN = eqjoinsel
 );
 
--- should use scalargesel and scalargejoinsel, but not supported in Postgres < 11
 CREATE OPERATOR >= (
 	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_ge,
 	COMMUTATOR = <= , NEGATOR = < ,
-	RESTRICT = scalargtsel, JOIN = scalargtjoinsel
+	RESTRICT = scalargesel, JOIN = scalargejoinsel
 );
 
 CREATE OPERATOR > (

src/hnsw.c

@@ -8,6 +8,7 @@
 #include "commands/progress.h"
 #include "commands/vacuum.h"
 #include "hnsw.h"
+#include "miscadmin.h"
 #include "utils/guc.h"
 #include "utils/selfuncs.h"
 
@@ -28,7 +29,7 @@ static relopt_kind hnsw_relopt_kind;
  * this grows bigger, we should use a shmem_request_hook and
  * RequestAddinShmemSpace() to pre-reserve space for this.
  */
-static void
+void
 HnswInitLockTranche(void)
 {
 	int		   *tranche_ids;
@@ -53,6 +54,7 @@ HnswInitLockTranche(void)
 void
 HnswInit(void)
 {
+	if (!process_shared_preload_libraries_in_progress)
 		HnswInitLockTranche();
 
 	hnsw_relopt_kind = add_reloption_kind();

src/hnsw.h

@@ -60,6 +60,7 @@
 #define PROGRESS_HNSW_PHASE_LOAD		2
 
 #define HNSW_MAX_SIZE (BLCKSZ - MAXALIGN(SizeOfPageHeaderData) - MAXALIGN(sizeof(HnswPageOpaqueData)) - sizeof(ItemIdData))
+#define HNSW_TUPLE_ALLOC_SIZE BLCKSZ
 
 #define HNSW_ELEMENT_TUPLE_SIZE(size)	MAXALIGN(offsetof(HnswElementTupleData, data) + (size))
 #define HNSW_NEIGHBOR_TUPLE_SIZE(level, m)	MAXALIGN(offsetof(HnswNeighborTupleData, indextids) + ((level) + 2) * (m) * sizeof(ItemPointerData))
@@ -79,7 +80,7 @@
 
 #if PG_VERSION_NUM < 130000
 #define list_delete_last(list) list_truncate(list, list_length(list) - 1)
-#define list_sort(list, cmp) list_qsort(list, cmp)
+#define list_sort(list, cmp) ((list) = list_qsort(list, cmp))
 #endif
 
 #define HnswIsElementTuple(tup) ((tup)->type == HNSW_ELEMENT_TUPLE_TYPE)
@@ -128,7 +129,7 @@ HnswPtrDeclare(HnswNeighborArray, HnswNeighborArrayRelptr, HnswNeighborArrayPtr)
 HnswPtrDeclare(HnswNeighborArrayPtr, HnswNeighborsRelptr, HnswNeighborsPtr);
 HnswPtrDeclare(char, DatumRelptr, DatumPtr);
 
-typedef struct HnswElementData
+struct HnswElementData
 {
 	HnswElementPtr next;
 	ItemPointerData heaptids[HNSW_HEAPTIDS];
@@ -143,7 +144,7 @@ typedef struct HnswElementData
 	BlockNumber neighborPage;
 	DatumPtr	value;
 	LWLock		lock;
-}			HnswElementData;
+};
 
 typedef HnswElementData * HnswElement;
 
@@ -154,12 +155,12 @@ typedef struct HnswCandidate
 	bool		closer;
 }			HnswCandidate;
 
-typedef struct HnswNeighborArray
+struct HnswNeighborArray
 {
 	int			length;
 	bool		closerSet;
 	HnswCandidate items[FLEXIBLE_ARRAY_MEMBER];
-}			HnswNeighborArray;
+};
 
 typedef struct HnswPairingHeapNode
 {
@@ -184,6 +185,7 @@ typedef struct HnswGraph
 
 	/* Entry state */
 	LWLock		entryLock;
+	LWLock		entryWaitLock;
 	HnswElementPtr entryPoint;
 
 	/* Allocations state */
@@ -354,6 +356,7 @@ typedef struct HnswVacuumState
 	/* Variables */
 	struct tidhash_hash *deleted;
 	BufferAccessStrategy bas;
+	HnswNeighborTuple ntup;
 	HnswElementData highestPoint;
 
 	/* Memory */
@@ -387,6 +390,7 @@ void		HnswLoadElement(HnswElement element, float *distance, Datum *q, Relation i
 void		HnswSetElementTuple(char *base, HnswElementTuple etup, HnswElement element);
 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);
 PGDLLEXPORT void HnswParallelBuildMain(dsm_segment *seg, shm_toc *toc);
 
 /* Index access methods */

src/hnswbuild.c

@@ -148,7 +148,6 @@ CreateGraphPages(HnswBuildState * buildstate)
 {
 	Relation	index = buildstate->index;
 	ForkNumber	forkNum = buildstate->forkNum;
-	Size		etupAllocSize;
 	Size		maxSize;
 	HnswElementTuple etup;
 	HnswNeighborTuple ntup;
@@ -160,12 +159,11 @@ CreateGraphPages(HnswBuildState * buildstate)
 	char	   *base = buildstate->hnswarea;
 
 	/* Calculate sizes */
-	etupAllocSize = BLCKSZ;
 	maxSize = HNSW_MAX_SIZE;
 
 	/* Allocate once */
-	etup = palloc0(etupAllocSize);
+	etup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
-	ntup = palloc0(BLCKSZ);
+	ntup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
 
 	/* Prepare first page */
 	buf = HnswNewBuffer(index, forkNum);
@@ -178,13 +176,13 @@ CreateGraphPages(HnswBuildState * buildstate)
 		Size		etupSize;
 		Size		ntupSize;
 		Size		combinedSize;
-		void	   *valuePtr = HnswPtrAccess(base, element->value);
+		Pointer		valuePtr = HnswPtrAccess(base, element->value);
 
 		/* Update iterator */
 		iter = element->next;
 
 		/* Zero memory for each element */
-		MemSet(etup, 0, etupAllocSize);
+		MemSet(etup, 0, HNSW_TUPLE_ALLOC_SIZE);
 
 		/* Calculate sizes */
 		etupSize = HNSW_ELEMENT_TUPLE_SIZE(VARSIZE_ANY(valuePtr));
@@ -192,7 +190,7 @@ CreateGraphPages(HnswBuildState * buildstate)
 		combinedSize = etupSize + ntupSize + sizeof(ItemIdData);
 
 		/* Initial size check */
-		if (etupSize > etupAllocSize)
+		if (etupSize > HNSW_TUPLE_ALLOC_SIZE)
 			elog(ERROR, "index tuple too large");
 
 		HnswSetElementTuple(base, etup, element);
@@ -254,38 +252,43 @@ WriteNeighborTuples(HnswBuildState * buildstate)
 	int			m = buildstate->m;
 	HnswElementPtr iter = buildstate->graph->head;
 	char	   *base = buildstate->hnswarea;
+	HnswNeighborTuple ntup;
+
+	/* Allocate once */
+	ntup = palloc0(HNSW_TUPLE_ALLOC_SIZE);
 
 	while (!HnswPtrIsNull(base, iter))
 	{
-		HnswElement e = HnswPtrAccess(base, iter);
+		HnswElement element = HnswPtrAccess(base, iter);
 		Buffer		buf;
 		Page		page;
-		ItemId		itemid;
+		Size		ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m);
-		HnswNeighborTuple ntup;
 
 		/* Update iterator */
-		iter = e->next;
+		iter = element->next;
+
+		/* Zero memory for each element */
+		MemSet(ntup, 0, HNSW_TUPLE_ALLOC_SIZE);
 
 		/* Can take a while, so ensure we can interrupt */
 		/* Needs to be called when no buffer locks are held */
 		CHECK_FOR_INTERRUPTS();
 
-		buf = ReadBufferExtended(index, forkNum, e->neighborPage, RBM_NORMAL, NULL);
+		buf = ReadBufferExtended(index, forkNum, element->neighborPage, RBM_NORMAL, NULL);
 		LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
 		page = BufferGetPage(buf);
-		itemid = PageGetItemId(page, e->neighborOffno);
-		ntup = (HnswNeighborTuple) PageGetItem(page, itemid);
 
-		/* Check expected size */
+		HnswSetNeighborTuple(base, ntup, element, m);
-		Assert(ItemIdGetLength(itemid) == HNSW_NEIGHBOR_TUPLE_SIZE(e->level, m));
 
-		/* Update page in-place */
+		if (!PageIndexTupleOverwrite(page, element->neighborOffno, (Item) ntup, ntupSize))
-		HnswSetNeighborTuple(base, ntup, e, m);
+			elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
 
 		/* Commit */
 		MarkBufferDirty(buf);
 		UnlockReleaseBuffer(buf);
 	}
+
+	pfree(ntup);
 }
 
 /*
@@ -428,10 +431,15 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
 	HnswGraph  *graph = buildstate->graph;
 	HnswElement entryPoint;
 	LWLock	   *entryLock = &graph->entryLock;
+	LWLock	   *entryWaitLock = &graph->entryWaitLock;
 	int			efConstruction = buildstate->efConstruction;
 	int			m = buildstate->m;
 	char	   *base = buildstate->hnswarea;
 
+	/* Wait if another process needs exclusive lock on entry lock */
+	LWLockAcquire(entryWaitLock, LW_EXCLUSIVE);
+	LWLockRelease(entryWaitLock);
+
 	/* Get entry point */
 	LWLockAcquire(entryLock, LW_SHARED);
 	entryPoint = HnswPtrAccess(base, graph->entryPoint);
@@ -442,8 +450,10 @@ InsertTupleInMemory(HnswBuildState * buildstate, HnswElement element)
 		/* Release shared lock */
 		LWLockRelease(entryLock);
 
-		/* Get exclusive lock */
+		/* Tell other processes to wait and get exclusive lock */
+		LWLockAcquire(entryWaitLock, LW_EXCLUSIVE);
 		LWLockAcquire(entryLock, LW_EXCLUSIVE);
+		LWLockRelease(entryWaitLock);
 
 		/* Get latest entry point after lock is acquired */
 		entryPoint = HnswPtrAccess(base, graph->entryPoint);
@@ -598,6 +608,9 @@ BuildCallback(Relation index, CALLBACK_ITEM_POINTER, Datum *values,
 static void
 InitGraph(HnswGraph * graph, char *base, long memoryTotal)
 {
+	/* Initialize the lock tranche if needed */
+	HnswInitLockTranche();
+
 	HnswPtrStore(base, graph->head, (HnswElement) NULL);
 	HnswPtrStore(base, graph->entryPoint, (HnswElement) NULL);
 	graph->memoryUsed = 0;
@@ -606,6 +619,7 @@ InitGraph(HnswGraph * graph, char *base, long memoryTotal)
 	graph->indtuples = 0;
 	SpinLockInit(&graph->lock);
 	LWLockInitialize(&graph->entryLock, hnsw_lock_tranche_id);
+	LWLockInitialize(&graph->entryWaitLock, hnsw_lock_tranche_id);
 	LWLockInitialize(&graph->allocatorLock, hnsw_lock_tranche_id);
 	LWLockInitialize(&graph->flushLock, hnsw_lock_tranche_id);
 }
@@ -971,6 +985,14 @@ HnswBeginParallel(HnswBuildState * buildstate, bool isconcurrent, int request)
 	/* Report less than allocated so never fails */
 	InitGraph(&hnswshared->graphData, hnswarea, esthnswarea - 1024 * 1024);
 
+	/*
+	 * Avoid base address for relptr for Postgres < 14.5
+	 * https://github.com/postgres/postgres/commit/7201cd18627afc64850537806da7f22150d1a83b
+	 */
+#if PG_VERSION_NUM < 140005
+	hnswshared->graphData.memoryUsed += MAXALIGN(1);
+#endif
+
 	shm_toc_insert(pcxt->toc, PARALLEL_KEY_HNSW_SHARED, hnswshared);
 	shm_toc_insert(pcxt->toc, PARALLEL_KEY_HNSW_AREA, hnswarea);
 

src/hnswinsert.c

@@ -355,9 +355,7 @@ HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, Hns
 			Buffer		buf;
 			Page		page;
 			GenericXLogState *state;
-			ItemId		itemid;
 			HnswNeighborTuple ntup;
-			Size		ntupSize;
 			int			idx = -1;
 			int			startIdx;
 			HnswElement neighborElement = HnswPtrAccess(base, hc->element);
@@ -396,9 +394,7 @@ HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, Hns
 			}
 
 			/* Get tuple */
-			itemid = PageGetItemId(page, offno);
+			ntup = (HnswNeighborTuple) PageGetItem(page, PageGetItemId(page, offno));
-			ntup = (HnswNeighborTuple) PageGetItem(page, itemid);
-			ntupSize = ItemIdGetLength(itemid);
 
 			/* Calculate index for update */
 			startIdx = (neighborElement->level - lc) * m;
@@ -427,13 +423,9 @@ HnswUpdateNeighborsOnDisk(Relation index, FmgrInfo *procinfo, Oid collation, Hns
 			{
 				ItemPointer indextid = &ntup->indextids[idx];
 
-				/* Update neighbor */
+				/* Update neighbor on the buffer */
 				ItemPointerSet(indextid, e->blkno, e->offno);
 
-				/* Overwrite tuple */
-				if (!PageIndexTupleOverwrite(page, offno, (Item) ntup, ntupSize))
-					elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
 				/* Commit */
 				if (building)
 					MarkBufferDirty(buf);
@@ -457,9 +449,7 @@ AddDuplicateOnDisk(Relation index, HnswElement element, HnswElement dup, bool bu
 	Buffer		buf;
 	Page		page;
 	GenericXLogState *state;
-	ItemId		itemid;
 	HnswElementTuple etup;
-	Size		etupSize;
 	int			i;
 
 	/* Read page */
@@ -477,9 +467,7 @@ AddDuplicateOnDisk(Relation index, HnswElement element, HnswElement dup, bool bu
 	}
 
 	/* Find space */
-	itemid = PageGetItemId(page, dup->offno);
+	etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, dup->offno));
-	etup = (HnswElementTuple) PageGetItem(page, itemid);
-	etupSize = ItemIdGetLength(itemid);
 	for (i = 0; i < HNSW_HEAPTIDS; i++)
 	{
 		if (!ItemPointerIsValid(&etup->heaptids[i]))
@@ -495,13 +483,9 @@ AddDuplicateOnDisk(Relation index, HnswElement element, HnswElement dup, bool bu
 		return false;
 	}
 
-	/* Add heap TID */
+	/* Add heap TID, modifying the tuple on the page directly */
 	etup->heaptids[i] = element->heaptids[0];
 
-	/* Overwrite tuple */
-	if (!PageIndexTupleOverwrite(page, dup->offno, (Item) etup, etupSize))
-		elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
 	/* Commit */
 	if (building)
 		MarkBufferDirty(buf);

src/hnswutils.c

@@ -860,12 +860,15 @@ HnswSearchLayer(char *base, Datum q, List *ep, int ef, int lc, Relation index, F
 static int
 #if PG_VERSION_NUM >= 130000
 CompareCandidateDistances(const ListCell *a, const ListCell *b)
+{
+	HnswCandidate *hca = lfirst(a);
+	HnswCandidate *hcb = lfirst(b);
 #else
 CompareCandidateDistances(const void *a, const void *b)
-#endif
 {
-	HnswCandidate *hca = lfirst((ListCell *) a);
+	HnswCandidate *hca = lfirst(*(ListCell **) a);
-	HnswCandidate *hcb = lfirst((ListCell *) b);
+	HnswCandidate *hcb = lfirst(*(ListCell **) b);
+#endif
 
 	if (hca->distance < hcb->distance)
 		return 1;
@@ -888,12 +891,15 @@ CompareCandidateDistances(const void *a, const void *b)
 static int
 #if PG_VERSION_NUM >= 130000
 CompareCandidateDistancesOffset(const ListCell *a, const ListCell *b)
+{
+	HnswCandidate *hca = lfirst(a);
+	HnswCandidate *hcb = lfirst(b);
 #else
 CompareCandidateDistancesOffset(const void *a, const void *b)
-#endif
 {
-	HnswCandidate *hca = lfirst((ListCell *) a);
+	HnswCandidate *hca = lfirst(*(ListCell **) a);
-	HnswCandidate *hcb = lfirst((ListCell *) b);
+	HnswCandidate *hcb = lfirst(*(ListCell **) b);
+#endif
 
 	if (hca->distance < hcb->distance)
 		return 1;
@@ -952,7 +958,9 @@ SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid col
 {
 	List	   *r = NIL;
 	List	   *w = list_copy(c);
-	pairingheap *wd;
+	HnswCandidate **wd;
+	int			wdlen = 0;
+	int			wdoff = 0;
 	HnswNeighborArray *neighbors = HnswGetNeighbors(base, e2, lc);
 	bool		mustCalculate = !neighbors->closerSet;
 	List	   *added = NIL;
@@ -961,7 +969,7 @@ SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid col
 	if (list_length(w) <= lm)
 		return w;
 
-	wd = pairingheap_allocate(CompareNearestCandidates, NULL);
+	wd = palloc(sizeof(HnswCandidate *) * list_length(w));
 
 	/* Ensure order of candidates is deterministic for closer caching */
 	if (sortCandidates)
@@ -1027,21 +1035,21 @@ SelectNeighbors(char *base, List *c, int lm, int lc, FmgrInfo *procinfo, Oid col
 		if (e->closer)
 			r = lappend(r, e);
 		else
-			pairingheap_add(wd, &(CreatePairingHeapNode(e)->ph_node));
+			wd[wdlen++] = e;
 	}
 
 	/* Cached value can only be used in future if sorted deterministically */
 	neighbors->closerSet = sortCandidates;
 
 	/* Keep pruned connections */
-	while (!pairingheap_is_empty(wd) && list_length(r) < lm)
+	while (wdoff < wdlen && list_length(r) < lm)
-		r = lappend(r, ((HnswPairingHeapNode *) pairingheap_remove_first(wd))->inner);
+		r = lappend(r, wd[wdoff++]);
 
 	/* Return pruned for update connections */
 	if (pruned != NULL)
 	{
-		if (!pairingheap_is_empty(wd))
+		if (wdoff < wdlen)
-			*pruned = ((HnswPairingHeapNode *) pairingheap_first(wd))->inner;
+			*pruned = wd[wdoff];
 		else
 			*pruned = linitial(w);
 	}

src/hnswvacuum.c

@@ -60,8 +60,7 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 		/* Iterate over nodes */
 		for (offno = FirstOffsetNumber; offno <= maxoffno; offno = OffsetNumberNext(offno))
 		{
-			ItemId		itemid = PageGetItemId(page, offno);
+			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, offno));
-			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, itemid);
 			int			idx = 0;
 			bool		itemUpdated = false;
 
@@ -92,15 +91,10 @@ RemoveHeapTids(HnswVacuumState * vacuumstate)
 
 				if (itemUpdated)
 				{
-					Size		etupSize = ItemIdGetLength(itemid);
-
 					/* Mark rest as invalid */
 					for (int i = idx; i < HNSW_HEAPTIDS; i++)
 						ItemPointerSetInvalid(&etup->heaptids[i]);
 
-					if (!PageIndexTupleOverwrite(page, offno, (Item) etup, etupSize))
-						elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
-
 					updated = true;
 				}
 			}
@@ -198,8 +192,8 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
 	FmgrInfo   *procinfo = vacuumstate->procinfo;
 	Oid			collation = vacuumstate->collation;
 	BufferAccessStrategy bas = vacuumstate->bas;
-	ItemId		itemid;
+	HnswNeighborTuple ntup = vacuumstate->ntup;
-	HnswNeighborTuple ntup;
+	Size		ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m);
 	char	   *base = NULL;
 
 	/* Skip if element is entry point */
@@ -213,23 +207,23 @@ RepairGraphElement(HnswVacuumState * vacuumstate, HnswElement element, HnswEleme
 	/* Find neighbors for element, skipping itself */
 	HnswFindElementNeighbors(base, element, entryPoint, index, procinfo, collation, m, efConstruction, true);
 
+	/* Zero memory for each element */
+	MemSet(ntup, 0, HNSW_TUPLE_ALLOC_SIZE);
+
+	/* Update neighbor tuple */
+	/* Do this before getting page to minimize locking */
+	HnswSetNeighborTuple(base, ntup, element, m);
+
 	/* Get neighbor page */
 	buf = ReadBufferExtended(index, MAIN_FORKNUM, element->neighborPage, RBM_NORMAL, bas);
 	LockBuffer(buf, BUFFER_LOCK_EXCLUSIVE);
 	state = GenericXLogStart(index);
 	page = GenericXLogRegisterBuffer(state, buf, 0);
 
-	/* Get neighbor tuple */
+	/* Overwrite tuple */
-	itemid = PageGetItemId(page, element->neighborOffno);
+	if (!PageIndexTupleOverwrite(page, element->neighborOffno, (Item) ntup, ntupSize))
-	ntup = (HnswNeighborTuple) PageGetItem(page, itemid);
-
-	/* Check expected size */
-	if (ItemIdGetLength(itemid) != HNSW_NEIGHBOR_TUPLE_SIZE(element->level, m))
 		elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
 
-	/* Update page in-place */
-	HnswSetNeighborTuple(base, ntup, element, m);
-
 	/* Commit */
 	GenericXLogFinish(state);
 	UnlockReleaseBuffer(buf);
@@ -482,11 +476,8 @@ MarkDeleted(HnswVacuumState * vacuumstate)
 		/* Update element and neighbors together */
 		for (offno = FirstOffsetNumber; offno <= maxoffno; offno = OffsetNumberNext(offno))
 		{
-			ItemId		itemid = PageGetItemId(page, offno);
+			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, PageGetItemId(page, offno));
-			HnswElementTuple etup = (HnswElementTuple) PageGetItem(page, itemid);
 			HnswNeighborTuple ntup;
-			Size		etupSize;
-			Size		ntupSize;
 			Buffer		nbuf;
 			Page		npage;
 			BlockNumber neighborPage;
@@ -510,10 +501,6 @@ MarkDeleted(HnswVacuumState * vacuumstate)
 			if (ItemPointerIsValid(&etup->heaptids[0]))
 				continue;
 
-			/* Calculate sizes */
-			etupSize = ItemIdGetLength(itemid);
-			ntupSize = HNSW_NEIGHBOR_TUPLE_SIZE(etup->level, vacuumstate->m);
-
 			/* Get neighbor page */
 			neighborPage = ItemPointerGetBlockNumber(&etup->neighbortid);
 			neighborOffno = ItemPointerGetOffsetNumber(&etup->neighbortid);
@@ -540,13 +527,10 @@ MarkDeleted(HnswVacuumState * vacuumstate)
 			for (int i = 0; i < ntup->count; i++)
 				ItemPointerSetInvalid(&ntup->indextids[i]);
 
-			/* Overwrite element tuple */
+			/*
-			if (!PageIndexTupleOverwrite(page, offno, (Item) etup, etupSize))
+			 * We modified the tuples in place, no need to call
-				elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
+			 * PageIndexTupleOverwrite
-
+			 */
-			/* Overwrite neighbor tuple */
-			if (!PageIndexTupleOverwrite(npage, neighborOffno, (Item) ntup, ntupSize))
-				elog(ERROR, "failed to add index item to \"%s\"", RelationGetRelationName(index));
 
 			/* Commit */
 			GenericXLogFinish(state);
@@ -591,6 +575,7 @@ InitVacuumState(HnswVacuumState * vacuumstate, IndexVacuumInfo *info, IndexBulkD
 	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);
@@ -610,6 +595,7 @@ FreeVacuumState(HnswVacuumState * vacuumstate)
 {
 	tidhash_destroy(vacuumstate->deleted);
 	FreeAccessStrategy(vacuumstate->bas);
+	pfree(vacuumstate->ntup);
 	MemoryContextDelete(vacuumstate->tmpCtx);
 }
 

src/vector.c

@@ -4,11 +4,18 @@
 
 #include "catalog/pg_type.h"
 #include "common/shortest_dec.h"
+#include "executor/execExpr.h"
 #include "fmgr.h"
 #include "hnsw.h"
 #include "ivfflat.h"
 #include "lib/stringinfo.h"
 #include "libpq/pqformat.h"
+#include "nodes/makefuncs.h"
+#include "nodes/nodeFuncs.h"
+#include "nodes/subscripting.h"
+#include "parser/parse_coerce.h"
+#include "parser/parse_expr.h"
+#include "parser/parse_node.h"
 #include "port.h"				/* for strtof() */
 #include "utils/array.h"
 #include "utils/builtins.h"
@@ -418,6 +425,180 @@ vector_send(PG_FUNCTION_ARGS)
 	PG_RETURN_BYTEA_P(pq_endtypsend(&buf));
 }
 
+/*
+ * Transform the subscript expressions
+ */
+static void
+vector_subscript_transform(SubscriptingRef *sbsref, List *indirection, ParseState *pstate, bool isSlice, bool isAssignment)
+{
+	A_Indices  *ai;
+	Node	   *subexpr;
+
+	if (list_length(indirection) != 1)
+		ereport(ERROR,
+				(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
+				 errmsg("vector allows only one subscript"),
+				 parser_errposition(pstate,
+									exprLocation((Node *) indirection))));
+
+	ai = linitial_node(A_Indices, indirection);
+
+	if (isSlice)
+	{
+		if (ai->lidx)
+		{
+			subexpr = transformExpr(pstate, ai->lidx, pstate->p_expr_kind);
+			/* If it's not int4 already, try to coerce */
+			subexpr = coerce_to_target_type(pstate,
+											subexpr, exprType(subexpr),
+											INT4OID, -1,
+											COERCION_ASSIGNMENT,
+											COERCE_IMPLICIT_CAST,
+											-1);
+			if (subexpr == NULL)
+				ereport(ERROR,
+						(errcode(ERRCODE_DATATYPE_MISMATCH),
+						 errmsg("vector subscript must have type integer"),
+						 parser_errposition(pstate, exprLocation(ai->lidx))));
+		}
+		else if (!ai->is_slice)
+		{
+			/* Make a constant 1 */
+			subexpr = (Node *) makeConst(INT4OID,
+										 -1,
+										 InvalidOid,
+										 sizeof(int32),
+										 Int32GetDatum(1),
+										 false,
+										 true); /* pass by value */
+		}
+		else
+		{
+			/* Slice with omitted lower bound, put NULL into the list */
+			subexpr = NULL;
+		}
+		sbsref->reflowerindexpr = list_make1(subexpr);
+	}
+	else
+		Assert(ai->lidx == NULL && !ai->is_slice);
+
+	if (ai->uidx)
+	{
+		subexpr = transformExpr(pstate, ai->uidx, pstate->p_expr_kind);
+		/* If it's not int4 already, try to coerce */
+		subexpr = coerce_to_target_type(pstate,
+										subexpr, exprType(subexpr),
+										INT4OID, -1,
+										COERCION_ASSIGNMENT,
+										COERCE_IMPLICIT_CAST,
+										-1);
+		if (subexpr == NULL)
+			ereport(ERROR,
+					(errcode(ERRCODE_DATATYPE_MISMATCH),
+					 errmsg("array subscript must have type integer"),
+					 parser_errposition(pstate, exprLocation(ai->uidx))));
+	}
+	else
+	{
+		/* Slice with omitted upper bound, put NULL into the list */
+		Assert(isSlice && ai->is_slice);
+		subexpr = NULL;
+	}
+	sbsref->refupperindexpr = list_make1(subexpr);
+
+	if (isSlice)
+		sbsref->refrestype = sbsref->refcontainertype;
+	else
+		sbsref->refrestype = FLOAT4OID;
+}
+
+/*
+ * Fetch a vector element
+ */
+static void
+vector_subscript_fetch(ExprState *state, ExprEvalStep *op, ExprContext *econtext)
+{
+	SubscriptingRefState *sbsrefstate = op->d.sbsref.state;
+	Vector	   *vec = DatumGetVector(*op->resvalue);
+	int			index = DatumGetInt32(sbsrefstate->upperindex[0]);
+
+	if (index < 1 || index > vec->dim)
+		*op->resnull = true;
+	else
+		*op->resvalue = Float4GetDatum(vec->x[index - 1]);
+}
+
+/*
+ * Fetch a vector slice
+ */
+static void
+vector_subscript_fetch_slice(ExprState *state, ExprEvalStep *op, ExprContext *econtext)
+{
+	SubscriptingRefState *sbsrefstate = op->d.sbsref.state;
+
+	if (sbsrefstate->upperprovided[0] && sbsrefstate->upperindexnull[0])
+		*op->resnull = true;
+	else if (sbsrefstate->lowerprovided[0] && sbsrefstate->lowerindexnull[0])
+		*op->resnull = true;
+	else
+	{
+		Vector	   *vec = DatumGetVector(*op->resvalue);
+		int			upperindex = sbsrefstate->upperprovided[0] ? DatumGetInt32(sbsrefstate->upperindex[0]) : vec->dim;
+		int			lowerindex = sbsrefstate->lowerprovided[0] ? DatumGetInt32(sbsrefstate->lowerindex[0]) : 1;
+		int			dim;
+		Vector	   *result;
+
+		if (upperindex > vec->dim)
+			upperindex = vec->dim;
+
+		if (lowerindex < 1)
+			lowerindex = 1;
+
+		dim = upperindex - lowerindex + 1;
+
+		CheckDim(dim);
+
+		result = InitVector(dim);
+		for (int i = 0; i < dim; i++)
+			result->x[i] = vec->x[lowerindex + i - 1];
+
+		*op->resvalue = PointerGetDatum(result);
+	}
+}
+
+/*
+ * Set up execution state for a vector subscript operation
+ */
+static void
+vector_exec_setup(const SubscriptingRef *sbsref, SubscriptingRefState *sbsrefstate, SubscriptExecSteps *methods)
+{
+	methods->sbs_check_subscripts = NULL;
+	if (sbsrefstate->numlower != 0)
+		methods->sbs_fetch = vector_subscript_fetch_slice;
+	else
+		methods->sbs_fetch = vector_subscript_fetch;
+	methods->sbs_assign = NULL;
+	methods->sbs_fetch_old = NULL;
+}
+
+/*
+ * Subscript handler
+ */
+PGDLLEXPORT PG_FUNCTION_INFO_V1(vector_subscript);
+Datum
+vector_subscript(PG_FUNCTION_ARGS)
+{
+	static const SubscriptRoutines sbsroutines = {
+		.transform = vector_subscript_transform,
+		.exec_setup = vector_exec_setup,
+		.fetch_strict = true,	/* fetch returns NULL for NULL inputs */
+		.fetch_leakproof = true,	/* fetch returns NULL for bad subscript */
+		.store_leakproof = false	/* ... but assignment throws error */
+	};
+
+	PG_RETURN_POINTER(&sbsroutines);
+}
+
 /*
  * Convert vector to vector
  * This is needed to check the type modifier
@@ -866,9 +1047,10 @@ vector_mul(PG_FUNCTION_ARGS)
 int
 vector_cmp_internal(Vector * a, Vector * b)
 {
-	CheckDims(a, b);
+	int			dim = Min(a->dim, b->dim);
 
-	for (int i = 0; i < a->dim; i++)
+	/* 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;
@@ -876,6 +1058,13 @@ vector_cmp_internal(Vector * a, Vector * b)
 		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;
 }
 
@@ -889,6 +1078,9 @@ vector_lt(PG_FUNCTION_ARGS)
 	Vector	   *a = PG_GETARG_VECTOR_P(0);
 	Vector	   *b = PG_GETARG_VECTOR_P(1);
 
+	/* TODO Remove in 0.7.0 */
+	CheckDims(a, b);
+
 	PG_RETURN_BOOL(vector_cmp_internal(a, b) < 0);
 }
 
@@ -902,6 +1094,9 @@ vector_le(PG_FUNCTION_ARGS)
 	Vector	   *a = PG_GETARG_VECTOR_P(0);
 	Vector	   *b = PG_GETARG_VECTOR_P(1);
 
+	/* TODO Remove in 0.7.0 */
+	CheckDims(a, b);
+
 	PG_RETURN_BOOL(vector_cmp_internal(a, b) <= 0);
 }
 
@@ -915,6 +1110,9 @@ vector_eq(PG_FUNCTION_ARGS)
 	Vector	   *a = PG_GETARG_VECTOR_P(0);
 	Vector	   *b = PG_GETARG_VECTOR_P(1);
 
+	/* TODO Remove in 0.7.0 */
+	CheckDims(a, b);
+
 	PG_RETURN_BOOL(vector_cmp_internal(a, b) == 0);
 }
 
@@ -928,6 +1126,9 @@ vector_ne(PG_FUNCTION_ARGS)
 	Vector	   *a = PG_GETARG_VECTOR_P(0);
 	Vector	   *b = PG_GETARG_VECTOR_P(1);
 
+	/* TODO Remove in 0.7.0 */
+	CheckDims(a, b);
+
 	PG_RETURN_BOOL(vector_cmp_internal(a, b) != 0);
 }
 
@@ -941,6 +1142,9 @@ vector_ge(PG_FUNCTION_ARGS)
 	Vector	   *a = PG_GETARG_VECTOR_P(0);
 	Vector	   *b = PG_GETARG_VECTOR_P(1);
 
+	/* TODO Remove in 0.7.0 */
+	CheckDims(a, b);
+
 	PG_RETURN_BOOL(vector_cmp_internal(a, b) >= 0);
 }
 
@@ -954,6 +1158,9 @@ vector_gt(PG_FUNCTION_ARGS)
 	Vector	   *a = PG_GETARG_VECTOR_P(0);
 	Vector	   *b = PG_GETARG_VECTOR_P(1);
 
+	/* TODO Remove in 0.7.0 */
+	CheckDims(a, b);
+
 	PG_RETURN_BOOL(vector_cmp_internal(a, b) > 0);
 }
 

test/expected/functions.out

@@ -24,6 +24,162 @@ SELECT '[1e37]'::vector * '[1e37]';
 ERROR:  value out of range: overflow
 SELECT '[1e-37]'::vector * '[1e-37]';
 ERROR:  value out of range: underflow
+SELECT ('[1,2,3]'::vector)[0];
+ vector 
+--------
+       
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[1];
+ vector 
+--------
+      1
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[2];
+ vector 
+--------
+      2
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[3];
+ vector 
+--------
+      3
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[4];
+ vector 
+--------
+       
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[1:1];
+ vector 
+--------
+ [1]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[1:2];
+ vector 
+--------
+ [1,2]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[2:4];
+ vector 
+--------
+ [2,3]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[-2:2];
+ vector 
+--------
+ [1,2]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[2:1];
+ERROR:  vector must have at least 1 dimension
+SELECT ('[1,2,3]'::vector)[:];
+ vector  
+---------
+ [1,2,3]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[:2];
+ vector 
+--------
+ [1,2]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[2:];
+ vector 
+--------
+ [2,3]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[:4];
+ vector  
+---------
+ [1,2,3]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[-2:];
+ vector  
+---------
+ [1,2,3]
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[NULL];
+ vector 
+--------
+       
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[NULL:2];
+ vector 
+--------
+ 
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[2:NULL];
+ vector 
+--------
+ 
+(1 row)
+
+SELECT ('[1,2,3]'::vector)[1][1];
+ERROR:  vector allows only one subscript
+SELECT '[1,2,3]'::vector = '[1,2,3]';
+ ?column? 
+----------
+ t
+(1 row)
+
+SELECT '[1,2,3]'::vector = '[1,2]';
+ERROR:  different vector dimensions 3 and 2
+SELECT vector_cmp('[1,2,3]', '[1,2,3]');
+ vector_cmp 
+------------
+          0
+(1 row)
+
+SELECT vector_cmp('[1,2,3]', '[0,0,0]');
+ vector_cmp 
+------------
+          1
+(1 row)
+
+SELECT vector_cmp('[0,0,0]', '[1,2,3]');
+ vector_cmp 
+------------
+         -1
+(1 row)
+
+SELECT vector_cmp('[1,2]', '[1,2,3]');
+ vector_cmp 
+------------
+         -1
+(1 row)
+
+SELECT vector_cmp('[1,2,3]', '[1,2]');
+ vector_cmp 
+------------
+          1
+(1 row)
+
+SELECT vector_cmp('[1,2]', '[2,3,4]');
+ vector_cmp 
+------------
+         -1
+(1 row)
+
+SELECT vector_cmp('[2,3]', '[1,2,3]');
+ vector_cmp 
+------------
+          1
+(1 row)
+
 SELECT vector_dims('[1,2,3]');
  vector_dims 
 -------------

test/sql/functions.sql

@@ -6,6 +6,37 @@ SELECT '[1,2,3]'::vector * '[4,5,6]';
 SELECT '[1e37]'::vector * '[1e37]';
 SELECT '[1e-37]'::vector * '[1e-37]';
 
+SELECT ('[1,2,3]'::vector)[0];
+SELECT ('[1,2,3]'::vector)[1];
+SELECT ('[1,2,3]'::vector)[2];
+SELECT ('[1,2,3]'::vector)[3];
+SELECT ('[1,2,3]'::vector)[4];
+SELECT ('[1,2,3]'::vector)[1:1];
+SELECT ('[1,2,3]'::vector)[1:2];
+SELECT ('[1,2,3]'::vector)[2:4];
+SELECT ('[1,2,3]'::vector)[-2:2];
+SELECT ('[1,2,3]'::vector)[2:1];
+SELECT ('[1,2,3]'::vector)[:];
+SELECT ('[1,2,3]'::vector)[:2];
+SELECT ('[1,2,3]'::vector)[2:];
+SELECT ('[1,2,3]'::vector)[:4];
+SELECT ('[1,2,3]'::vector)[-2:];
+SELECT ('[1,2,3]'::vector)[NULL];
+SELECT ('[1,2,3]'::vector)[NULL:2];
+SELECT ('[1,2,3]'::vector)[2:NULL];
+SELECT ('[1,2,3]'::vector)[1][1];
+
+SELECT '[1,2,3]'::vector = '[1,2,3]';
+SELECT '[1,2,3]'::vector = '[1,2]';
+
+SELECT vector_cmp('[1,2,3]', '[1,2,3]');
+SELECT vector_cmp('[1,2,3]', '[0,0,0]');
+SELECT vector_cmp('[0,0,0]', '[1,2,3]');
+SELECT vector_cmp('[1,2]', '[1,2,3]');
+SELECT vector_cmp('[1,2,3]', '[1,2]');
+SELECT vector_cmp('[1,2]', '[2,3,4]');
+SELECT vector_cmp('[2,3]', '[1,2,3]');
+
 SELECT vector_dims('[1,2,3]');
 
 SELECT round(vector_norm('[1,1]')::numeric, 5);

vector.control

@@ -1,4 +1,4 @@
 comment = 'vector data type and ivfflat and hnsw access methods'
-default_version = '0.5.1'
+default_version = '0.6.1'
 module_pathname = '$libdir/vector'
 relocatable = true