Added concatenate operator for vectors

CHANGELOG.md

@@ -1,15 +1,15 @@
 ## 0.7.0 (unreleased)
 
-- Added `subvector` function
+- Added concatenate operator for vectors
 
-## 0.6.2 (2024-03-18)
+## 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 segmentation fault with `shared_preload_libraries`
+- Fixed error with `shared_preload_libraries`
 - Fixed vector subtraction being marked as commutative
 
 ## 0.6.0 (2024-01-29)

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": "0.6.2",
+	"version": "0.6.1",
 	"maintainer": [
 		"Andrew Kane <andrew@ankane.org>"
 	],
@@ -20,7 +20,7 @@
 		"vector": {
 			"file": "sql/vector.sql",
 			"docfile": "README.md",
-			"version": "0.6.2",
+			"version": "0.6.1",
 			"abstract": "Open-source vector similarity search for Postgres"
 		}
 	},

Makefile

@@ -1,5 +1,5 @@
 EXTENSION = vector
-EXTVERSION = 0.6.2
+EXTVERSION = 0.6.1
 
 MODULE_big = vector
 DATA = $(wildcard sql/*--*.sql)

Makefile.win

@@ -1,5 +1,5 @@
 EXTENSION = vector
-EXTVERSION = 0.6.2
+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

@@ -20,7 +20,7 @@ Compile and install the extension (supports Postgres 12+)
 
 ```sh
 cd /tmp
-git clone --branch v0.6.2 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
@@ -45,7 +45,7 @@ Then use `nmake` to build:
 ```cmd
 set "PGROOT=C:\Program Files\PostgreSQL\16"
 cd %TEMP%
-git clone --branch v0.6.2 https://github.com/pgvector/pgvector.git
+git clone --branch v0.6.1 https://github.com/pgvector/pgvector.git
 cd pgvector
 nmake /F Makefile.win
 nmake /F Makefile.win install
@@ -604,18 +604,6 @@ and query with:
 SELECT * FROM items ORDER BY embedding::vector(3) <-> '[3,1,2]' LIMIT 5;
 ```
 
-#### Are binary vectors supported?
-
-You can store binary vectors and perform exact nearest neighbor search by Hamming distance in Postgres without an extension ([example](https://github.com/pgvector/pgvector-python/blob/master/examples/hash_image_search.py)).
-
-```tsql
-CREATE TABLE items (id bigserial PRIMARY KEY, embedding bit(3));
-INSERT INTO items (embedding) VALUES (B'000'), (B'111');
-SELECT * FROM items ORDER BY bit_count(embedding # B'101') LIMIT 5;
-```
-
-Indexing is not currently supported.
-
 #### Do indexes need to fit into memory?
 
 No, but like other index types, you’ll likely see better performance if they do. You can get the size of an index with:
@@ -694,6 +682,7 @@ Operator | Description | Added
 \+ | element-wise addition |
 \- | element-wise subtraction |
 \* | element-wise multiplication | 0.5.0
+\|\| | concatenate | 0.7.0
 <-> | Euclidean distance |
 <#> | negative inner product |
 <=> | cosine distance |
@@ -706,7 +695,6 @@ cosine_distance(vector, vector) → double precision | cosine distance |
 inner_product(vector, vector) → double precision | inner product |
 l2_distance(vector, vector) → double precision | Euclidean distance |
 l1_distance(vector, vector) → double precision | taxicab distance | 0.5.0
-subvector(vector, integer, integer) → vector | subvector | 0.7.0 [unreleased]
 vector_dims(vector) → integer | number of dimensions |
 vector_norm(vector) → double precision | Euclidean norm |
 
@@ -792,7 +780,7 @@ This adds pgvector to the [Postgres image](https://hub.docker.com/_/postgres) (r
 You can also build the image manually:
 
 ```sh
-git clone --branch v0.6.2 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=16 -t myuser/pgvector .
 ```

sql/vector--0.6.1--0.6.2.sql

@@ -1,2 +0,0 @@
--- complain if script is sourced in psql, rather than via CREATE EXTENSION
-\echo Use "ALTER EXTENSION vector UPDATE TO '0.6.2'" to load this file. \quit

sql/vector--0.6.2--0.7.0.sql

@@ -1,5 +1,9 @@
 -- 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 subvector(vector, int, int) RETURNS vector
+CREATE FUNCTION vector_concat(vector, vector) RETURNS vector
 	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
+
+CREATE OPERATOR || (
+	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_concat
+);

sql/vector.sql

@@ -58,9 +58,6 @@ CREATE FUNCTION vector_sub(vector, vector) RETURNS vector
 CREATE FUNCTION vector_mul(vector, vector) RETURNS vector
 	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
 
-CREATE FUNCTION subvector(vector, int, int) RETURNS vector
-	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
-
 -- private functions
 
 CREATE FUNCTION vector_lt(vector, vector) RETURNS bool
@@ -102,6 +99,9 @@ CREATE FUNCTION vector_avg(double precision[]) RETURNS vector
 CREATE FUNCTION vector_combine(double precision[], double precision[]) RETURNS double precision[]
 	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
 
+CREATE FUNCTION vector_concat(vector, vector) RETURNS vector
+	AS 'MODULE_PATHNAME' LANGUAGE C IMMUTABLE STRICT PARALLEL SAFE;
+
 -- aggregates
 
 CREATE AGGREGATE avg(vector) (
@@ -227,6 +227,10 @@ CREATE OPERATOR > (
 	RESTRICT = scalargtsel, JOIN = scalargtjoinsel
 );
 
+CREATE OPERATOR || (
+	LEFTARG = vector, RIGHTARG = vector, PROCEDURE = vector_concat
+);
+
 -- access methods
 
 CREATE FUNCTION ivfflathandler(internal) RETURNS index_am_handler

src/vector.c

@@ -860,37 +860,6 @@ vector_mul(PG_FUNCTION_ARGS)
 	PG_RETURN_POINTER(result);
 }
 
-/*
- * Get a subset of a vector
- */
-PGDLLEXPORT PG_FUNCTION_INFO_V1(subvector);
-Datum
-subvector(PG_FUNCTION_ARGS)
-{
-	Vector	   *a = PG_GETARG_VECTOR_P(0);
-	int32		start = PG_GETARG_INT32(1);
-	int32		count = PG_GETARG_INT32(2);
-	int32		end = start + count;
-	float	   *ax = a->x;
-	Vector	   *result;
-	int			dim;
-
-	if (start < 1)
-		start = 1;
-
-	if (end > a->dim)
-		end = a->dim + 1;
-
-	dim = end - start;
-	CheckDim(dim);
-	result = InitVector(dim);
-
-	for (int i = 0; i < dim; i++)
-		result->x[i] = ax[start - 1 + i];
-
-	PG_RETURN_POINTER(result);
-}
-
 /*
  * Internal helper to compare vectors
  */
@@ -1191,3 +1160,28 @@ vector_avg(PG_FUNCTION_ARGS)
 
 	PG_RETURN_POINTER(result);
 }
+
+/*
+ * Concat vectors
+ */
+PGDLLEXPORT PG_FUNCTION_INFO_V1(vector_concat);
+Datum
+vector_concat(PG_FUNCTION_ARGS)
+{
+	Vector	   *a = PG_GETARG_VECTOR_P(0);
+	Vector	   *b = PG_GETARG_VECTOR_P(1);
+	Vector	   *result;
+	int			dim = a->dim + b->dim;
+
+	CheckDim(dim);
+
+	result = InitVector(dim);
+
+	for (int i = 0; i < a->dim; i++)
+		result->x[i] = a->x[i];
+
+	for (int i = 0; i < b->dim; i++)
+		result->x[i + a->dim] = b->x[i];
+
+	PG_RETURN_POINTER(result);
+}

test/expected/functions.out

@@ -24,6 +24,14 @@ 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 || '[4,5]'::vector;
+  ?column?   
+-------------
+ [1,2,3,4,5]
+(1 row)
+
+SELECT array_fill(0, ARRAY[16000])::vector || '[1]'::vector;
+ERROR:  vector cannot have more than 16000 dimensions
 SELECT '[1,2,3]'::vector = '[1,2,3]';
  ?column? 
 ----------
@@ -208,34 +216,6 @@ SELECT l1_distance('[3e38]', '[-3e38]');
     Infinity
 (1 row)
 
-SELECT subvector('[1,2,3,4,5]', 1, 3);
- subvector 
------------
- [1,2,3]
-(1 row)
-
-SELECT subvector('[1,2,3,4,5]', 3, 2);
- subvector 
------------
- [3,4]
-(1 row)
-
-SELECT subvector('[1,2,3,4,5]', -1, 3);
- subvector 
------------
- [1]
-(1 row)
-
-SELECT subvector('[1,2,3,4,5]', 3, 9);
- subvector 
------------
- [3,4,5]
-(1 row)
-
-SELECT subvector('[1,2,3,4,5]', 1, 0);
-ERROR:  vector must have at least 1 dimension
-SELECT subvector('[1,2,3,4,5]', -1, 2);
-ERROR:  vector must have at least 1 dimension
 SELECT avg(v) FROM unnest(ARRAY['[1,2,3]'::vector, '[3,5,7]']) v;
     avg    
 -----------

test/expected/input.out

@@ -116,30 +116,8 @@ SELECT '[1, ,3]'::vector;
 ERROR:  invalid input syntax for type vector: "[1, ,3]"
 LINE 1: SELECT '[1, ,3]'::vector;
                ^
-SELECT '[1,2,3]'::vector(3);
- vector  
----------
- [1,2,3]
-(1 row)
-
 SELECT '[1,2,3]'::vector(2);
 ERROR:  expected 2 dimensions, not 3
-SELECT '[1,2,3]'::vector(3, 2);
-ERROR:  invalid type modifier
-LINE 1: SELECT '[1,2,3]'::vector(3, 2);
-                          ^
-SELECT '[1,2,3]'::vector('a');
-ERROR:  invalid input syntax for type integer: "a"
-LINE 1: SELECT '[1,2,3]'::vector('a');
-                          ^
-SELECT '[1,2,3]'::vector(0);
-ERROR:  dimensions for type vector must be at least 1
-LINE 1: SELECT '[1,2,3]'::vector(0);
-                          ^
-SELECT '[1,2,3]'::vector(16001);
-ERROR:  dimensions for type vector cannot exceed 16000
-LINE 1: SELECT '[1,2,3]'::vector(16001);
-                          ^
 SELECT unnest('{"[1,2,3]", "[4,5,6]"}'::vector[]);
  unnest  
 ---------

test/sql/functions.sql

@@ -6,6 +6,9 @@ SELECT '[1,2,3]'::vector * '[4,5,6]';
 SELECT '[1e37]'::vector * '[1e37]';
 SELECT '[1e-37]'::vector * '[1e-37]';
 
+SELECT '[1,2,3]'::vector || '[4,5]'::vector;
+SELECT array_fill(0, ARRAY[16000])::vector || '[1]'::vector;
+
 SELECT '[1,2,3]'::vector = '[1,2,3]';
 SELECT '[1,2,3]'::vector = '[1,2]';
 
@@ -48,13 +51,6 @@ SELECT l1_distance('[0,0]', '[0,1]');
 SELECT l1_distance('[1,2]', '[3]');
 SELECT l1_distance('[3e38]', '[-3e38]');
 
-SELECT subvector('[1,2,3,4,5]', 1, 3);
-SELECT subvector('[1,2,3,4,5]', 3, 2);
-SELECT subvector('[1,2,3,4,5]', -1, 3);
-SELECT subvector('[1,2,3,4,5]', 3, 9);
-SELECT subvector('[1,2,3,4,5]', 1, 0);
-SELECT subvector('[1,2,3,4,5]', -1, 2);
-
 SELECT avg(v) FROM unnest(ARRAY['[1,2,3]'::vector, '[3,5,7]']) v;
 SELECT avg(v) FROM unnest(ARRAY['[1,2,3]'::vector, '[3,5,7]', NULL]) v;
 SELECT avg(v) FROM unnest(ARRAY[]::vector[]) v;

test/sql/input.sql

@@ -22,13 +22,7 @@ SELECT '[1,]'::vector;
 SELECT '[1a]'::vector;
 SELECT '[1,,3]'::vector;
 SELECT '[1, ,3]'::vector;
-
-SELECT '[1,2,3]'::vector(3);
 SELECT '[1,2,3]'::vector(2);
-SELECT '[1,2,3]'::vector(3, 2);
-SELECT '[1,2,3]'::vector('a');
-SELECT '[1,2,3]'::vector(0);
-SELECT '[1,2,3]'::vector(16001);
 
 SELECT unnest('{"[1,2,3]", "[4,5,6]"}'::vector[]);
 SELECT '{"[1,2,3]"}'::vector(2)[];

vector.control

@@ -1,4 +1,4 @@
 comment = 'vector data type and ivfflat and hnsw access methods'
-default_version = '0.6.2'
+default_version = '0.6.1'
 module_pathname = '$libdir/vector'
 relocatable = true