Repository files navigation

Open-source vector similarity search for Postgres

Store your vectors with the rest of your data. Supports:

• exact and approximate nearest neighbor search
• single-precision, half-precision, binary, and sparse vectors
• L2 distance, inner product, cosine distance, L1 distance, Hamming distance, and Jaccard distance
• anylanguage with a Postgres client

PlusACID compliance, point-in-time recovery, JOINs, and all of the othergreat features of Postgres

Compile and install the extension (supports Postgres 13+)

cd /tmpgit clone --branch v0.8.0 https://github.com/pgvector/pgvector.gitcd pgvectormakemake install# may need sudo

See theinstallation notes if you run into issues

You can also install it withDocker,Homebrew,PGXN,APT,Yum,pkg, orconda-forge, and it comes preinstalled withPostgres.app and manyhosted providers. There are also instructions forGitHub Actions.

EnsureC++ support in Visual Studio is installed, and run:

call"C:\Program Files\Microsoft Visual Studio\2022\Community\VC\Auxiliary\Build\vcvars64.bat"

Note: The exact path will vary depending on your Visual Studio version and edition

Then usenmake to build:

set"PGROOT=C:\Program Files\PostgreSQL\16"cd%TEMP%git clone --branch v0.8.0 https://github.com/pgvector/pgvector.gitcd pgvectornmake /F Makefile.winnmake /F Makefile.win install

Note: Postgres 17 is not supported with MSVC yet due to anupstream issue

See theinstallation notes if you run into issues

You can also install it withDocker orconda-forge.

Enable the extension (do this once in each database where you want to use it)

CREATE EXTENSION vector;

Create a vector column with 3 dimensions

CREATETABLEitems (idbigserialPRIMARY KEY, embedding vector(3));

Insert vectors

INSERT INTO items (embedding)VALUES ('[1,2,3]'), ('[4,5,6]');

Get the nearest neighbors by L2 distance

SELECT*FROM itemsORDER BY embedding<->'[3,1,2]'LIMIT5;

Also supports inner product (<#>), cosine distance (<=>), and L1 distance (<+>, added in 0.7.0)

Note:<#> returns the negative inner product since Postgres only supportsASC order index scans on operators

Create a new table with a vector column

CREATETABLEitems (idbigserialPRIMARY KEY, embedding vector(3));

Or add a vector column to an existing table

ALTERTABLE items ADD COLUMN embedding vector(3);

Also supportshalf-precision,binary, andsparse vectors

Insert vectors

INSERT INTO items (embedding)VALUES ('[1,2,3]'), ('[4,5,6]');

Or load vectors in bulk usingCOPY (example)

COPY items (embedding)FROM STDIN WITH (FORMAT BINARY);

Upsert vectors

INSERT INTO items (id, embedding)VALUES (1,'[1,2,3]'), (2,'[4,5,6]')ON CONFLICT (id) DOUPDATESET embedding=EXCLUDED.embedding;

Update vectors

UPDATE itemsSET embedding='[1,2,3]'WHERE id=1;

Delete vectors

DELETEFROM itemsWHERE id=1;

Get the nearest neighbors to a vector

SELECT*FROM itemsORDER BY embedding<->'[3,1,2]'LIMIT5;

Supported distance functions are:

• <-> - L2 distance
• <#> - (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

SELECT*FROM itemsWHERE id!=1ORDER BY embedding<-> (SELECT embeddingFROM itemsWHERE id=1)LIMIT5;

Get rows within a certain distance

SELECT*FROM itemsWHERE embedding<->'[3,1,2]'<5;

Note: Combine withORDER BY andLIMIT to use an index

Get the distance

SELECT embedding<->'[3,1,2]'AS distanceFROM items;

For inner product, multiply by -1 (since<#> returns the negative inner product)

SELECT (embedding<#>'[3,1,2]')*-1AS inner_productFROM items;

For cosine similarity, use 1 - cosine distance

SELECT1- (embedding<=>'[3,1,2]')AS cosine_similarityFROM items;

Average vectors

SELECTAVG(embedding)FROM items;

Average groups of vectors

SELECT category_id,AVG(embedding)FROM itemsGROUP BY category_id;

By default, pgvector performs exact nearest neighbor search, which provides perfect recall.

You can add an index to use approximate nearest neighbor search, which trades some recall for speed. Unlike typical indexes, you will see different results for queries after adding an approximate index.

Supported index types are:

• HNSW
• IVFFlat

An HNSW index creates a multilayer graph. It has better query performance than IVFFlat (in terms of speed-recall tradeoff), but has slower build times and uses more memory. Also, an index can be created without any data in the table since there isn’t a training step like IVFFlat.

Add an index for each distance function you want to use.

L2 distance

CREATEINDEXON items USING hnsw (embedding vector_l2_ops);

Note: Usehalfvec_l2_ops forhalfvec andsparsevec_l2_ops forsparsevec (and similar with the other distance functions)

Inner product

CREATEINDEXON items USING hnsw (embedding vector_ip_ops);

Cosine distance

CREATEINDEXON items USING hnsw (embedding vector_cosine_ops);

L1 distance - added in 0.7.0

CREATEINDEXON items USING hnsw (embedding vector_l1_ops);

Hamming distance - added in 0.7.0

CREATEINDEXON items USING hnsw (embedding bit_hamming_ops);

Jaccard distance - added in 0.7.0

CREATEINDEXON items USING hnsw (embedding bit_jaccard_ops);

Supported types are:

• vector - up to 2,000 dimensions
• halfvec - up to 4,000 dimensions (added in 0.7.0)
• bit - up to 64,000 dimensions (added in 0.7.0)
• sparsevec - up to 1,000 non-zero elements (added in 0.7.0)

Specify HNSW parameters

• m - the max number of connections per layer (16 by default)
• ef_construction - the size of the dynamic candidate list for constructing the graph (64 by default)

CREATEINDEXON items USING hnsw (embedding vector_l2_ops) WITH (m=16, ef_construction=64);

A higher value ofef_construction provides better recall at the cost of index build time / insert speed.

Specify the size of the dynamic candidate list for search (40 by default)

SEThnsw.ef_search=100;

A higher value provides better recall at the cost of speed.

UseSET LOCAL inside a transaction to set it for a single query

BEGIN;SET LOCALhnsw.ef_search=100;SELECT ...COMMIT;

Indexes build significantly faster when the graph fits intomaintenance_work_mem

SET maintenance_work_mem='8GB';

A notice is shown when the graph no longer fits

NOTICE:  hnsw graph no longer fits into maintenance_work_mem after 100000 tuplesDETAIL:  Building will take significantly more time.HINT:  Increase maintenance_work_mem to speed up builds.

Note: Do not setmaintenance_work_mem so high that it exhausts the memory on the server

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)

SET max_parallel_maintenance_workers=7;-- plus leader

For a large number of workers, you may also need to increasemax_parallel_workers (8 by default)

Checkindexing progress

SELECT phase, round(100.0* blocks_done/ nullif(blocks_total,0),1)AS"%"FROM pg_stat_progress_create_index;

The phases for HNSW are:

1. initializing
2. loading tuples

An IVFFlat index divides vectors into lists, and then searches a subset of those lists that are closest to the query vector. It has faster build times and uses less memory than HNSW, but has lower query performance (in terms of speed-recall tradeoff).

Three keys to achieving good recall are:

1. Create the indexafter the table has some data
2. Choose an appropriate number of lists - a good place to start isrows / 1000 for up to 1M rows andsqrt(rows) for over 1M rows
3. When querying, specify an appropriate number ofprobes (higher is better for recall, lower is better for speed) - a good place to start issqrt(lists)

Add an index for each distance function you want to use.

L2 distance

CREATEINDEXON items USING ivfflat (embedding vector_l2_ops) WITH (lists=100);

Note: Usehalfvec_l2_ops forhalfvec (and similar with the other distance functions)

Inner product

CREATEINDEXON items USING ivfflat (embedding vector_ip_ops) WITH (lists=100);

Cosine distance

CREATEINDEXON items USING ivfflat (embedding vector_cosine_ops) WITH (lists=100);

Hamming distance - added in 0.7.0

CREATEINDEXON items USING ivfflat (embedding bit_hamming_ops) WITH (lists=100);

Supported types are:

• vector - up to 2,000 dimensions
• halfvec - up to 4,000 dimensions (added in 0.7.0)
• bit - up to 64,000 dimensions (added in 0.7.0)

Specify the number of probes (1 by default)

SETivfflat.probes=10;

A higher value provides better recall at the cost of speed, and it can be set to the number of lists for exact nearest neighbor search (at which point the planner won’t use the index)

UseSET LOCAL inside a transaction to set it for a single query

BEGIN;SET LOCALivfflat.probes=10;SELECT ...COMMIT;

Speed up index creation on large tables by increasing the number of parallel workers (2 by default)

SET max_parallel_maintenance_workers=7;-- plus leader

For a large number of workers, you may also need to increasemax_parallel_workers (8 by default)

Checkindexing progress

SELECT phase, round(100.0* tuples_done/ nullif(tuples_total,0),1)AS"%"FROM pg_stat_progress_create_index;

The phases for IVFFlat are:

1. initializing
2. performing k-means
3. assigning tuples
4. loading tuples

Note:% is only populated during theloading tuples phase

There are a few ways to index nearest neighbor queries with aWHERE clause.

SELECT*FROM itemsWHERE category_id=123ORDER BY embedding<->'[3,1,2]'LIMIT5;

A good place to start is creating an index on the filter column. This can provide fast, exact nearest neighbor search in many cases. Postgres has a number ofindex types for this: B-tree (default), hash, GiST, SP-GiST, GIN, and BRIN.

CREATEINDEXON items (category_id);

For multiple columns, consider amulticolumn index.

CREATEINDEXON items (location_id, category_id);

Exact indexes work well for conditions that match a low percentage of rows. Otherwise,approximate indexes can work better.

CREATEINDEXON items USING hnsw (embedding vector_l2_ops);

With approximate indexes, filtering is appliedafter the index is scanned. If a condition matches 10% of rows, with HNSW and the defaulthnsw.ef_search of 40, only 4 rows will match on average. For more rows, increasehnsw.ef_search.

SEThnsw.ef_search=200;

Starting with 0.8.0, you can enableiterative index scans, which will automatically scan more of the index when needed.

SEThnsw.iterative_scan= strict_order;

If filtering by only a few distinct values, considerpartial indexing.

CREATEINDEXON items USING hnsw (embedding vector_l2_ops)WHERE (category_id=123);

If filtering by many different values, considerpartitioning.

CREATETABLEitems (embedding vector(3), category_idint) PARTITION BY LIST(category_id);

Added in 0.8.0

With approximate indexes, queries with filtering can return less results since filtering is appliedafter the index is scanned. Starting with 0.8.0, you can enable iterative index scans, which will automatically scan more of the index until enough results are found (or it reacheshnsw.max_scan_tuples orivfflat.max_probes).

Iterative scans can use strict or relaxed ordering.

Strict ensures results are in the exact order by distance

SEThnsw.iterative_scan= strict_order;

Relaxed allows results to be slightly out of order by distance, but provides better recall

SEThnsw.iterative_scan= relaxed_order;# orSETivfflat.iterative_scan= relaxed_order;

With relaxed ordering, you can use amaterialized CTE to get strict ordering

WITH relaxed_resultsAS MATERIALIZED (SELECT id, embedding<->'[1,2,3]'AS distanceFROM itemsWHERE category_id=123ORDER BY distanceLIMIT5)SELECT*FROM relaxed_resultsORDER BY distance;

For queries that filter by distance, use a materialized CTE and place the distance filter outside of it for best performance (due to thecurrent behavior of the Postgres executor)

WITH nearest_resultsAS MATERIALIZED (SELECT id, embedding<->'[1,2,3]'AS distanceFROM itemsORDER BY distanceLIMIT5)SELECT*FROM nearest_resultsWHERE distance<5ORDER BY distance;

Note: Place any other filters inside the CTE

Since scanning a large portion of an approximate index is expensive, there are options to control when a scan ends.

Specify the max number of tuples to visit (20,000 by default)

SEThnsw.max_scan_tuples=20000;

Note: This is approximate and does not affect the initial scan

Specify the max amount of memory to use, as a multiple ofwork_mem (1 by default)

SEThnsw.scan_mem_multiplier=2;

Note: Try increasing this if increasinghnsw.max_scan_tuples does not improve recall

Specify the max number of probes

SETivfflat.max_probes=100;

Note: If this is lower thanivfflat.probes,ivfflat.probes will be used

Added in 0.7.0

Use thehalfvec type to store half-precision vectors

CREATETABLEitems (idbigserialPRIMARY KEY, embedding halfvec(3));

Added in 0.7.0

Index vectors at half precision for smaller indexes

CREATEINDEXON items USING hnsw ((embedding::halfvec(3)) halfvec_l2_ops);

Get the nearest neighbors

SELECT*FROM itemsORDER BY embedding::halfvec(3)<->'[1,2,3]'LIMIT5;

Use thebit type to store binary vectors (example)

CREATETABLEitems (idbigserialPRIMARY KEY, embeddingbit(3));INSERT INTO items (embedding)VALUES ('000'), ('111');

Get the nearest neighbors by Hamming distance (added in 0.7.0)

SELECT*FROM itemsORDER BY embedding<~>'101'LIMIT5;

Or (before 0.7.0)

SELECT*FROM itemsORDER BY bit_count(embedding# '101') LIMIT 5;

Also supports Jaccard distance (<%>)

Added in 0.7.0

Use expression indexing for binary quantization

CREATEINDEXON items USING hnsw ((binary_quantize(embedding)::bit(3)) bit_hamming_ops);

Get the nearest neighbors by Hamming distance

SELECT*FROM itemsORDER BY binary_quantize(embedding)::bit(3)<~> binary_quantize('[1,-2,3]')LIMIT5;

Re-rank by the original vectors for better recall

SELECT*FROM (SELECT*FROM itemsORDER BY binary_quantize(embedding)::bit(3)<~> binary_quantize('[1,-2,3]')LIMIT20)ORDER BY embedding<=>'[1,-2,3]'LIMIT5;

Added in 0.7.0

Use thesparsevec type to store sparse vectors

CREATETABLEitems (idbigserialPRIMARY KEY, embedding sparsevec(5));

Insert vectors

INSERT INTO items (embedding)VALUES ('{1:1,3:2,5:3}/5'), ('{1:4,3:5,5:6}/5');

The format is{index1:value1,index2:value2}/dimensions and indices start at 1 like SQL arrays

Get the nearest neighbors by L2 distance

SELECT*FROM itemsORDER BY embedding<->'{1:3,3:1,5:2}/5'LIMIT5;

Use together with Postgresfull-text search for hybrid search.

SELECT id, contentFROM items, plainto_tsquery('hello search') queryWHERE textsearch @@ queryORDER BY ts_rank_cd(textsearch, query)DESCLIMIT5;

You can useReciprocal Rank Fusion or across-encoder to combine results.

Added in 0.7.0

Use expression indexing to index subvectors

CREATEINDEXON items USING hnsw ((subvector(embedding,1,3)::vector(3)) vector_cosine_ops);

Get the nearest neighbors by cosine distance

SELECT*FROM itemsORDER BY subvector(embedding,1,3)::vector(3)<=> subvector('[1,2,3,4,5]'::vector,1,3)LIMIT5;

Re-rank by the full vectors for better recall

SELECT*FROM (SELECT*FROM itemsORDER BY subvector(embedding,1,3)::vector(3)<=> subvector('[1,2,3,4,5]'::vector,1,3)LIMIT20)ORDER BY embedding<=>'[1,2,3,4,5]'LIMIT5;

Use a tool likePgTune to set initial values for Postgres server parameters. For instance,shared_buffers should typically be 25% of the server’s memory. You can find the config file with:

SHOW config_file;

And check individual settings with:

SHOW shared_buffers;

Be sure to restart Postgres for changes to take effect.

UseCOPY for bulk loading data (example).

COPY items (embedding)FROM STDIN WITH (FORMAT BINARY);

Add any indexesafter loading the initial data for best performance.

See index build time forHNSW andIVFFlat.

In production environments, create indexes concurrently to avoid blocking writes.

CREATEINDEXCONCURRENTLY ...

UseEXPLAIN ANALYZE to debug performance.

EXPLAIN ANALYZESELECT*FROM itemsORDER BY embedding<->'[3,1,2]'LIMIT5;

To speed up queries without an index, increasemax_parallel_workers_per_gather.

SET max_parallel_workers_per_gather=4;

If vectors are normalized to length 1 (likeOpenAI embeddings), use inner product for best performance.

SELECT*FROM itemsORDER BY embedding<#>'[3,1,2]'LIMIT5;

To speed up queries with an IVFFlat index, increase the number of inverted lists (at the expense of recall).

CREATEINDEXON items USING ivfflat (embedding vector_l2_ops) WITH (lists=1000);

Vacuuming can take a while for HNSW indexes. Speed it up by reindexing first.

REINDEX INDEX CONCURRENTLY index_name;VACUUM table_name;

Monitor performance withpg_stat_statements (be sure to add it toshared_preload_libraries).

CREATE EXTENSION pg_stat_statements;

Get the most time-consuming queries with:

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_minFROM pg_stat_statementsORDER BY total_plan_time+ total_exec_timeDESCLIMIT20;

Note: Replacetotal_plan_time + total_exec_time withtotal_time for Postgres < 13

Monitor recall by comparing results from approximate search with exact search.

BEGIN;SET LOCAL enable_indexscan= off;-- use exact searchSELECT ...COMMIT;

Scale pgvector the same way you scale Postgres.

Scale vertically by increasing memory, CPU, and storage on a single instance. Use existing tools totune parameters andmonitor performance.

Scale horizontally withreplicas, or useCitus or another approach for sharding (example).

Use pgvector from any language with a Postgres client. You can even generate and store vectors in one language and query them in another.

A non-partitioned table has a limit of 32 TB by default in Postgres. A partitioned table can have thousands of partitions of that size.

Yes, pgvector uses the write-ahead log (WAL), which allows for replication and point-in-time recovery.

You can usehalf-precision indexing to index up to 4,000 dimensions orbinary quantization to index up to 64,000 dimensions. Another option isdimensionality reduction.

You can usevector as the type (instead ofvector(3)).

CREATETABLEembeddings (model_idbigint, item_idbigint, embedding vector,PRIMARY KEY (model_id, item_id));

However, you can only create indexes on rows with the same number of dimensions (usingexpression andpartial indexing):

CREATEINDEXON embeddings USING hnsw ((embedding::vector(3)) vector_l2_ops)WHERE (model_id=123);

and query with:

SELECT*FROM embeddingsWHERE model_id=123ORDER BY embedding::vector(3)<->'[3,1,2]'LIMIT5;

You can use thedouble precision[] ornumeric[] type to store vectors with more precision.

CREATETABLEitems (idbigserialPRIMARY KEY, embeddingdouble precision[]);-- use {} instead of [] for Postgres arraysINSERT INTO items (embedding)VALUES ('{1,2,3}'), ('{4,5,6}');

Optionally, add acheck constraint to ensure data can be converted to thevector type and has the expected dimensions.

ALTERTABLE items ADDCHECK (vector_dims(embedding::vector)=3);

Useexpression indexing to index (at a lower precision):

CREATEINDEXON items USING hnsw ((embedding::vector(3)) vector_l2_ops);

and query with:

SELECT*FROM itemsORDER BY embedding::vector(3)<->'[3,1,2]'LIMIT5;

No, but like other index types, you’ll likely see better performance if they do. You can get the size of an index with:

SELECT pg_size_pretty(pg_relation_size('index_name'));

The query needs to have anORDER BY andLIMIT, and theORDER BY must be the result of a distance operator (not an expression) in ascending order.

-- indexORDER BY embedding<=>'[3,1,2]'LIMIT5;-- no indexORDER BY1- (embedding<=>'[3,1,2]')DESCLIMIT5;

You can encourage the planner to use an index for a query with:

BEGIN;SET LOCAL enable_seqscan= off;SELECT ...COMMIT;

Also, if the table is small, a table scan may be faster.

The planner doesn’t considerout-of-line storage in cost estimates, which can make a serial scan look cheaper. You can reduce the cost of a parallel scan for a query with:

BEGIN;SET LOCAL min_parallel_table_scan_size=1;SET LOCAL parallel_setup_cost=1;SELECT ...COMMIT;

or choose to store vectors inline:

ALTERTABLE items ALTER COLUMN embeddingSET STORAGE PLAIN;

Results are limited by the size of the dynamic candidate list (hnsw.ef_search), which is 40 by default. There may be even less results due to dead tuples or filtering conditions in the query. Enablingiterative index scans can help address this.

Also, note thatNULL vectors are not indexed (as well as zero vectors for cosine distance).

The index was likely created with too little data for the number of lists. Drop the index until the table has more data.

DROPINDEX index_name;

Results can also be limited by the number of probes (ivfflat.probes). Enablingiterative index scans can address this.

Also, note thatNULL vectors are not indexed (as well as zero vectors for cosine distance).

• Vector
• Halfvec
• Bit
• Sparsevec

Each vector takes4 * dimensions + 8 bytes of storage. Each element is a single-precision floating-point number (like thereal type in Postgres), and all elements must be finite (noNaN,Infinity or-Infinity). Vectors can have up to 16,000 dimensions.

Each half vector takes2 * dimensions + 8 bytes of storage. Each element is a half-precision floating-point number, and all elements must be finite (noNaN,Infinity or-Infinity). Half vectors can have up to 16,000 dimensions.

Each bit vector takesdimensions / 8 + 8 bytes of storage. See thePostgres docs for more info.

Each sparse vector takes8 * non-zero elements + 16 bytes of storage. Each element is a single-precision floating-point number, and all elements must be finite (noNaN,Infinity or-Infinity). Sparse vectors can have up to 16,000 non-zero elements.

If your machine has multiple Postgres installations, specify the path topg_config with:

export PG_CONFIG=/Library/PostgreSQL/17/bin/pg_config

Then re-run the installation instructions (runmake clean beforemake if needed). Ifsudo is needed formake install, use:

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

Note: Replace17 with your Postgres server version

If compilation fails withfatal error: postgres.h: No such file or directory, make sure Postgres development files are installed on the server.

For Ubuntu and Debian, use:

sudo apt install postgresql-server-dev-17

Note: Replace17 with your Postgres server version

If compilation fails and the output includeswarning: no such sysroot directory on Mac, reinstall Xcode Command Line Tools.

By default, pgvector compiles with-march=native on some platforms for best performance. However, this can lead toIllegal instruction errors if trying to run the compiled extension on a different machine.

To compile for portability, use:

make OPTFLAGS=""

If compilation fails withCannot open include file: 'postgres.h': No such file or directory, make surePGROOT is correct.

If installation fails withAccess is denied, re-run the installation instructions as an administrator.

Get theDocker image with:

docker pull pgvector/pgvector:pg17

This adds pgvector to thePostgres image (replace17 with your Postgres server version, and run it the same way).

You can also build the image manually:

git clone --branch v0.8.0 https://github.com/pgvector/pgvector.gitcd pgvectordocker build --pull --build-arg PG_MAJOR=17 -t myuser/pgvector.

With Homebrew Postgres, you can use:

brew install pgvector

Note: This only adds it to thepostgresql@17 andpostgresql@14 formulas

Install from thePostgreSQL Extension Network with:

pgxn install vector

Debian and Ubuntu packages are available from thePostgreSQL APT Repository. Follow thesetup instructions and run:

sudo apt install postgresql-17-pgvector

Note: Replace17 with your Postgres server version

RPM packages are available from thePostgreSQL Yum Repository. Follow thesetup instructions for your distribution and run:

sudo yum install pgvector_17# orsudo dnf install pgvector_17

Note: Replace17 with your Postgres server version

Install the FreeBSD package with:

pkg install postgresql16-pgvector

or the port with:

cd /usr/ports/databases/pgvectormake install

With Conda Postgres, install fromconda-forge with:

conda install -c conda-forge pgvector

This method iscommunity-maintained by@mmcauliffe

Download thelatest release with Postgres 15+.

pgvector is available onthese providers.

Install the latest version (use the same method as the original installation). Then in each database you want to upgrade, run:

ALTER EXTENSION vectorUPDATE;

You can check the version in the current database with:

SELECT extversionFROM pg_extensionWHERE extname='vector';

If upgrading with Postgres 12, remove this line fromsql/vector--0.5.1--0.6.0.sql:

ALTERTYPE vectorSET (STORAGE= external);

Then runmake install andALTER EXTENSION vector UPDATE;.

The Docker image is now published in thepgvector org, and there are tags for each supported version of Postgres (rather than alatest tag).

docker pull pgvector/pgvector:pg16# ordocker pull pgvector/pgvector:0.6.0-pg16

Also, if you’ve increasedmaintenance_work_mem, make sure--shm-size is at least that size to avoid an error with parallel HNSW index builds.

docker run --shm-size=1g ...

Thanks to:

• PASE: PostgreSQL Ultra-High-Dimensional Approximate Nearest Neighbor Search Extension
• Faiss: A Library for Efficient Similarity Search and Clustering of Dense Vectors
• Using the Triangle Inequality to Accelerate k-means
• k-means++: The Advantage of Careful Seeding
• Concept Decompositions for Large Sparse Text Data using Clustering
• Efficient and Robust Approximate Nearest Neighbor Search using Hierarchical Navigable Small World Graphs

View thechangelog

Everyone is encouraged to help improve this project. Here are a few ways you can help:

• Report bugs
• Fix bugs andsubmit pull requests
• Write, clarify, or fix documentation
• Suggest or add new features

To get started with development:

git clone https://github.com/pgvector/pgvector.gitcd pgvectormakemake install

To run all tests:

make installcheck# regression testsmake prove_installcheck# TAP tests

To run single tests:

make installcheck REGRESS=functions# regression testmake prove_installcheck PROVE_TESTS=test/t/001_ivfflat_wal.pl# TAP test

To enable assertions:

make clean&& PG_CFLAGS="-DUSE_ASSERT_CHECKING" make&& make install

To enable benchmarking:

make clean&& PG_CFLAGS="-DIVFFLAT_BENCH" make&& make install

To show memory usage:

make clean&& PG_CFLAGS="-DHNSW_MEMORY -DIVFFLAT_MEMORY" make&& make install

To get k-means metrics:

make clean&& PG_CFLAGS="-DIVFFLAT_KMEANS_DEBUG" make&& make install

Resources for contributors

• Extension Building Infrastructure
• Index Access Method Interface Definition
• Generic WAL Records