PGEXT.CLOUD – CONTRIB

earthdistance

Mon, 01 Jan 0001 00:00:00 +0000

earthdistance : calculate great-circle distances on the surface of the Earth

Overview

ID	Extension	Package	Version	Category	License	Language
1690	earthdistance	earthdistance	`1.2`	GIS	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
Requires	cube
See Also	postgis q3c pg_sphere pointcloud h3 pg_geohash pg_eviltransform qdgc pgrouting pg_polyline

Packages

PG18	PG17	PG16	PG15	PG14
1.2	1.2	1.2	1.2	1.2

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION earthdistance CASCADE; -- requires cube

Usage

earthdistance: Great circle distances on the surface of the Earth

The earthdistance module provides two different approaches to calculating great circle distances on the surface of the Earth. The Earth is assumed to be perfectly spherical (for more accuracy, use PostGIS).

CREATE EXTENSION earthdistance CASCADE; -- requires cube

Cube-Based Earth Distances

Data is stored in cubes using 3 coordinates representing x, y, z distance from Earth’s center. A earth domain over type cube is provided.

Functions

Function	Description
`earth()` → `float8`	Returns the assumed radius of the Earth
`sec_to_gc(float8)` → `float8`	Converts straight line (secant) distance to great circle distance
`gc_to_sec(float8)` → `float8`	Converts great circle distance to straight line (secant) distance
`ll_to_earth(float8, float8)` → `earth`	Returns location given latitude and longitude in degrees
`latitude(earth)` → `float8`	Returns latitude in degrees
`longitude(earth)` → `float8`	Returns longitude in degrees
`earth_distance(earth, earth)` → `float8`	Returns great circle distance between two points
`earth_box(earth, float8)` → `cube`	Returns a box for indexed search using cube `@>` operator

Example

-- Distance between New York and London in meters
SELECT earth_distance(
  ll_to_earth(40.7128, -74.0060),
  ll_to_earth(51.5074, -0.1278)
);

-- Find all points within 1000 meters of a location (indexable)
SELECT *
FROM places
WHERE earth_box(ll_to_earth(40.7128, -74.0060), 1000) @> ll_to_earth(lat, lon);

Point-Based Earth Distances

Represents Earth locations as point type values where the first component is longitude and the second is latitude (in degrees).

Operators

Operator	Description
`point <@> point` → `float8`	Distance in statute miles between two points

Example

-- Distance in statute miles
SELECT point(-74.0060, 40.7128) <@> point(-0.1278, 51.5074);

Note: Units are hardwired to statute miles. The point-based approach has edge condition issues near poles and ±180° longitude; the cube-based representation avoids these discontinuities.

fuzzystrmatch

Mon, 01 Jan 0001 00:00:00 +0000

fuzzystrmatch : determine similarities and distance between strings

Overview

ID	Extension	Package	Version	Category	License	Language
2380	fuzzystrmatch	fuzzystrmatch	`1.2`	FTS	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	postgis_tiger_geocoder
See Also	pg_trgm pg_similarity smlar pgpcre re2 pg_bigm biscuit unaccent citext

Packages

PG18	PG17	PG16	PG15	PG14
1.2	1.2	1.2	1.2	1.2

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION fuzzystrmatch;

Usage

fuzzystrmatch: Determine similarities and distance between strings

The fuzzystrmatch module provides functions to determine similarities and distance between strings.

CREATE EXTENSION fuzzystrmatch;

Soundex

Converts a string to its Soundex code (useful for matching similar-sounding names):

SELECT soundex('Anne'), soundex('Ann'), difference('Anne', 'Ann');
-- A500, A500, 4

SELECT soundex('Anne'), soundex('Andrew'), difference('Anne', 'Andrew');
-- A500, A536, 2

SELECT soundex('Anne'), soundex('Margaret'), difference('Anne', 'Margaret');
-- A500, M626, 0

The difference function returns 0–4, where 4 means most similar.

Daitch-Mokotoff Soundex

Returns a set of Daitch-Mokotoff soundex codes (better for non-English names):

SELECT daitch_mokotoff('George');
-- {595000}

SELECT daitch_mokotoff('John');
-- {160000,460000}

-- Find names sounding like 'Schwartzenegger'
SELECT * FROM s WHERE daitch_mokotoff(nm) && daitch_mokotoff('Schwartzenegger');

Supports indexing with GIN:

CREATE INDEX ON s USING gin (daitch_mokotoff(nm) gin__int_ops);

Levenshtein Distance

Computes edit distance between two strings (insertions, deletions, substitutions):

SELECT levenshtein('GUMBO', 'GAMBOL');
-- 2

SELECT levenshtein('GUMBO', 'GAMBOL', 2, 1, 1);
-- 3 (custom costs: insert=2, delete=1, substitute=1)

-- Bounded version (faster, stops early)
SELECT levenshtein_less_equal('extensive', 'exhaustive', 2);
-- 3 (actual distance exceeds threshold, returns actual)

SELECT levenshtein_less_equal('extensive', 'exhaustive', 4);
-- 4

Metaphone

Returns a metaphone code for a string:

SELECT metaphone('GUMBO', 4);
-- KM

Double Metaphone

Returns primary and alternate codes (handles more name variations):

SELECT dmetaphone('gumbo');
-- KMP

SELECT dmetaphone_alt('gumbo');
-- KMP

pg_trgm

Mon, 01 Jan 0001 00:00:00 +0000

pg_trgm : text similarity measurement and index searching based on trigrams

Overview

ID	Extension	Package	Version	Category	License	Language
2390	pg_trgm	pg_trgm	`1.6`	FTS	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	pg_similarity fuzzystrmatch smlar pg_bigm pgpcre re2 biscuit citext unaccent pgroonga

Packages

PG18	PG17	PG16	PG15	PG14
1.6	1.6	1.6	1.6	1.6

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION pg_trgm;

Usage

pg_trgm: Text similarity measurement and index searching based on trigrams

The pg_trgm module provides functions and operators for determining the similarity of alphanumeric text based on trigram matching, plus index operator classes for fast string similarity searches.

CREATE EXTENSION pg_trgm;

A trigram is a group of three consecutive characters from a string. Two strings are similar if they share many trigrams.

Functions

Function	Description
`similarity(text, text)` → `real`	Returns similarity between 0 and 1
`show_trgm(text)` → `text[]`	Returns array of all trigrams in the string
`word_similarity(text, text)` → `real`	Similarity of first string to most similar word in second
`strict_word_similarity(text, text)` → `real`	Similar but with stricter word boundary matching
`show_limit()` → `real`	(Deprecated) Returns `pg_trgm.similarity_threshold`
`set_limit(real)` → `real`	(Deprecated) Sets `pg_trgm.similarity_threshold`

SELECT similarity('word', 'two words');
-- 0.36363637

SELECT show_trgm('word');
-- {"  w"," wo",ord,"rd ",wor}

Operators

Operator	Description
`text % text` → `boolean`	True if similarity > `pg_trgm.similarity_threshold`
`text <% text` → `boolean`	True if word similarity > `pg_trgm.word_similarity_threshold`
`text %> text` → `boolean`	Commutator of `<%`
`text <<% text` → `boolean`	True if strict word similarity > threshold
`text %>> text` → `boolean`	Commutator of `<<%`
`text <-> text` → `real`	Distance (1 - similarity)
`text <<-> text` → `real`	Word distance (1 - word_similarity)
`text <->> text` → `real`	Commutator of `<<->`
`text <<<-> text` → `real`	Strict word distance
`text <->>> text` → `real`	Commutator of `<<<->`

GUC Parameters

Parameter	Default	Description
`pg_trgm.similarity_threshold`	0.3	Threshold for `%` operator
`pg_trgm.word_similarity_threshold`	0.6	Threshold for `<%` and `%>` operators
`pg_trgm.strict_word_similarity_threshold`	0.5	Threshold for `<<%` and `%>>` operators

Index Support

GiST and GIN indexes support the similarity operators:

-- GIN index (faster lookups, slower builds)
CREATE INDEX trgm_idx ON test_trgm USING GIN (t gin_trgm_ops);

-- GiST index (supports distance operators for KNN)
CREATE INDEX trgm_idx ON test_trgm USING GIST (t gist_trgm_ops);

-- GiST with custom signature length
CREATE INDEX trgm_idx ON test_trgm USING GIST (t gist_trgm_ops(siglen=32));

Text Search Example

Using trigram indexes to speed up LIKE / ILIKE / regex queries:

SELECT t, similarity(t, 'word') AS sml
FROM test_trgm
WHERE t % 'word'
ORDER BY sml DESC, t;

-- KNN search using distance operator
SELECT t, t <-> 'word' AS dist
FROM test_trgm
ORDER BY dist
LIMIT 10;

GIN and GiST trigram indexes also accelerate LIKE, ILIKE, ~, and ~* queries automatically.

tablefunc

Mon, 01 Jan 0001 00:00:00 +0000

tablefunc : functions that manipulate whole tables, including crosstab

Overview

ID	Extension	Package	Version	Category	License	Language
2590	tablefunc	tablefunc	`1.0`	OLAP	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	pg4ml
See Also	pg_duckdb first_last_agg extra_window_functions argm pg_lake plprql duckdb_fdw

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION tablefunc;

Usage

tablefunc: Functions that manipulate whole tables, including crosstab

The tablefunc module provides functions that return tables (multiple rows). It is a trusted extension available to non-superusers with CREATE privilege.

CREATE EXTENSION tablefunc;

normal_rand – Generate Random Values

Produces a set of normally distributed random values (Gaussian distribution):

SELECT * FROM normal_rand(1000, 5, 3);
-- numvals: number of values, mean: 5, stddev: 3

crosstab(text) – Single-Parameter Pivot

Pivots data from long format to wide format. The SQL must return row_name, category, and value columns:

CREATE TABLE ct(id SERIAL, rowid TEXT, attribute TEXT, value TEXT);
INSERT INTO ct(rowid, attribute, value) VALUES
  ('test1','att1','val1'), ('test1','att2','val2'),
  ('test1','att3','val3'), ('test2','att1','val5'),
  ('test2','att2','val6'), ('test2','att3','val7');

SELECT *
FROM crosstab(
  'SELECT rowid, attribute, value FROM ct ORDER BY 1,2'
) AS ct(row_name text, category_1 text, category_2 text, category_3 text);

 row_name | category_1 | category_2 | category_3
----------+------------+------------+------------
 test1    | val1       | val2       | val3
 test2    | val5       | val6       | val7

The input query should always use ORDER BY 1,2 to ensure proper grouping. Extra output columns beyond the available values are filled with nulls.

crosstab(text, text) – Two-Parameter Pivot with Categories

Handles cases where some groups may not have data for all categories:

CREATE TABLE sales(year int, month int, qty int);
INSERT INTO sales VALUES(2007,1,1000),(2007,2,1500),(2007,7,500),
  (2007,11,1500),(2007,12,2000),(2008,1,1000);

SELECT * FROM crosstab(
  'SELECT year, month, qty FROM sales ORDER BY 1',
  'SELECT m FROM generate_series(1,12) m'
) AS (
  year int, "Jan" int, "Feb" int, "Mar" int, "Apr" int,
  "May" int, "Jun" int, "Jul" int, "Aug" int,
  "Sep" int, "Oct" int, "Nov" int, "Dec" int
);

 year | Jan  | Feb  | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov  | Dec
------+------+------+-----+-----+-----+-----+-----+-----+-----+-----+------+------
 2007 | 1000 | 1500 |     |     |     |     | 500 |     |     |     | 1500 | 2000
 2008 | 1000 |      |     |     |     |     |     |     |     |     |      |

The source SQL may include “extra” columns between row_name and category/value.

crosstab2, crosstab3, crosstab4 – Predefined Wrappers

Pre-built wrappers that avoid writing the FROM clause (text input/output only):

SELECT * FROM crosstab3(
  'SELECT rowid, attribute, value FROM ct ORDER BY 1,2'
);

Create custom wrappers for other types:

CREATE TYPE my_crosstab_float8_5_cols AS (
    my_row_name text, c1 float8, c2 float8, c3 float8, c4 float8, c5 float8
);

CREATE OR REPLACE FUNCTION crosstab_float8_5_cols(text)
    RETURNS setof my_crosstab_float8_5_cols
    AS '$libdir/tablefunc','crosstab' LANGUAGE C STABLE STRICT;

connectby – Hierarchical Tree Display

Displays hierarchical data stored in a table with key and parent-key fields:

CREATE TABLE connectby_tree(keyid text, parent_keyid text, pos int);
INSERT INTO connectby_tree VALUES
  ('row1',NULL,0), ('row2','row1',0), ('row3','row1',0),
  ('row4','row2',1), ('row5','row2',0), ('row6','row4',0),
  ('row7','row3',0), ('row8','row6',0), ('row9','row5',0);

-- With branch display and ordering
SELECT * FROM connectby('connectby_tree', 'keyid', 'parent_keyid', 'pos', 'row2', 0, '~')
  AS t(keyid text, parent_keyid text, level int, branch text, pos int);

 keyid | parent_keyid | level |       branch        | pos
-------+--------------+-------+---------------------+-----
 row2  |              |     0 | row2                |   1
 row5  | row2         |     1 | row2~row5           |   2
 row9  | row5         |     2 | row2~row5~row9      |   3
 row4  | row2         |     1 | row2~row4           |   4
 row6  | row4         |     2 | row2~row4~row6      |   5
 row8  | row6         |     3 | row2~row4~row6~row8 |   6

Parameters: connectby(relname, keyid_fld, parent_keyid_fld [, orderby_fld], start_with, max_depth [, branch_delim])

start_with: key value of the row to start at (text)
max_depth: maximum depth to descend (0 = unlimited)
branch_delim: separator for branch output (optional, default ~)
Index the parent-key field for performance on large tables

bloom

Mon, 01 Jan 0001 00:00:00 +0000

bloom : bloom access method - signature file based index

Overview

ID	Extension	Package	Version	Category	License	Language
2990	bloom	bloom	`1.0`	FEAT	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	rum hll roaringbitmap pgfaceting parray_gin btree_gin btree_gist intarray

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION bloom;

Usage

bloom: bloom access method - signature file based index

The bloom extension provides an index access method based on Bloom filters. A Bloom filter is a space-efficient data structure that tests whether an element is a member of a set, with possible false positives but no false negatives.

Bloom indexes are particularly useful for tables with many columns where queries test arbitrary combinations of columns. A single bloom index can replace multiple btree indexes while using significantly less space.

Create Bloom Index

CREATE INDEX bloomidx ON tbloom USING bloom (i1, i2, i3, i4, i5, i6);

With custom parameters:

CREATE INDEX bloomidx ON tbloom USING bloom (i1, i2, i3)
       WITH (length=80, col1=2, col2=2, col3=4);

Parameters

Parameter	Default	Max	Description
`length`	80	4096	Length of each signature in bits (rounded to nearest 16)
`col1` - `col32`	2	4095	Number of bits generated for each index column

Operator Class Support

Bloom indexes only support the equality operator (=). Built-in operator classes exist for int4 and text. You can define custom operator classes for other types that have a hash function:

CREATE OPERATOR CLASS text_ops
DEFAULT FOR TYPE text USING bloom AS
    OPERATOR    1   =(text, text),
    FUNCTION    1   hashtext(text);

Example

-- Create a table with many columns
CREATE TABLE tbloom AS
   SELECT
     (random() * 1000000)::int as i1,
     (random() * 1000000)::int as i2,
     (random() * 1000000)::int as i3,
     (random() * 1000000)::int as i4,
     (random() * 1000000)::int as i5,
     (random() * 1000000)::int as i6
   FROM generate_series(1, 10000000);

-- A single bloom index covers all column combinations
CREATE INDEX bloomidx ON tbloom USING bloom (i1, i2, i3, i4, i5, i6);

-- Queries on any subset of columns can use the index
SELECT * FROM tbloom WHERE i2 = 898732 AND i5 = 123451;

Limitations

Only equality (=) queries are supported (no range queries)
Does not support UNIQUE indexes
Does not support searching for NULL values
Results require recheck against heap due to false positives

pltcl

Mon, 01 Jan 0001 00:00:00 +0000

pltcl : PL/Tcl procedural language

Overview

ID	Extension	Package	Version	Category	License	Language
3240	pltcl	pltcl	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Schemas	`pg_catalog`
See Also	plperl plperlu plpython3u pljava plsh plr plxslt plpgsql
Siblings	pltclu

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION pltcl;

Usage

pltcl: PL/Tcl trusted procedural language

PL/Tcl allows writing PostgreSQL functions in the Tcl programming language. As a trusted language, it restricts access to the filesystem and other system resources.

CREATE EXTENSION pltcl;

-- Simple function returning a greeting
CREATE FUNCTION tcl_hello(name text) RETURNS text
LANGUAGE pltcl AS $$
  return "Hello, $1!"
$$;

SELECT tcl_hello('world');

-- Function with conditional logic
CREATE FUNCTION tcl_max(a integer, b integer) RETURNS integer
LANGUAGE pltcl AS $$
  if {$1 > $2} {return $1}
  return $2
$$;

-- Set-returning function
CREATE FUNCTION tcl_sequence(count integer) RETURNS SETOF integer
LANGUAGE pltcl AS $$
  for {set i 1} {$i <= $1} {incr i} {
    return_next $i
  }
$$;

-- Trigger function
CREATE FUNCTION tcl_audit() RETURNS trigger
LANGUAGE pltcl AS $$
  set NEW(modified_at) [clock format [clock seconds] -format "%Y-%m-%d %H:%M:%S"]
  return [array get NEW]
$$;

Database access from PL/Tcl uses the spi_exec command:

CREATE FUNCTION tcl_count_rows(tbl text) RETURNS integer
LANGUAGE pltcl AS $$
  spi_exec "SELECT count(*) AS cnt FROM $1"
  return $cnt
$$;

pltclu

Mon, 01 Jan 0001 00:00:00 +0000

pltcl : PL/TclU untrusted procedural language

Overview

ID	Extension	Package	Version	Category	License	Language
3250	pltclu	pltcl	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d--	No	No	No	Yes	no	no

Relationships
Schemas	`pg_catalog`
See Also	plperl plperlu plpython3u pljava plsh plr plxslt plpgsql
Siblings	pltcl

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION pltclu;

Usage

pltclu: PL/Tcl untrusted procedural language

PL/Tcl Untrusted provides full Tcl capabilities including filesystem access and external program execution. Only superusers can create functions in this language.

CREATE EXTENSION pltclu;

-- Read a file from the server filesystem
CREATE FUNCTION read_file(filename text) RETURNS text
LANGUAGE pltclu AS $$
  set fd [open $1 r]
  set content [read $fd]
  close $fd
  return $content
$$;

-- Execute an external command
CREATE FUNCTION run_command(cmd text) RETURNS text
LANGUAGE pltclu AS $$
  return [exec {*}$1]
$$;

-- Access environment variables
CREATE FUNCTION get_env(varname text) RETURNS text
LANGUAGE pltclu AS $$
  if {[info exists ::env($1)]} {
    return $::env($1)
  }
  return ""
$$;

SELECT get_env('HOME');

plperl

Mon, 01 Jan 0001 00:00:00 +0000

plperl : PL/Perl procedural language

Overview

ID	Extension	Package	Version	Category	License	Language
3260	plperl	plperl	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Schemas	`pg_catalog`
Need By	bool_plperl hstore_plperl jsonb_plperl sparql
See Also	plperlu plpython3u pllua pltcl plsh pljava plr plxslt plpgsql plv8
Siblings	bool_plperl hstore_plperl jsonb_plperl

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION plperl;

Usage

plperl: PL/Perl trusted procedural language

PL/Perl allows writing PostgreSQL functions in Perl. As a trusted language, it runs in a restricted environment without access to the filesystem or external modules.

CREATE EXTENSION plperl;

-- Simple scalar function
CREATE FUNCTION perl_hello(text) RETURNS text
LANGUAGE plperl AS $$
  my ($name) = @_;
  return "Hello, $name!";
$$;

SELECT perl_hello('world');

-- Using Perl regex for text processing
CREATE FUNCTION clean_whitespace(text) RETURNS text
LANGUAGE plperl AS $$
  my ($str) = @_;
  $str =~ s/^\s+|\s+$//g;   # trim
  $str =~ s/\s+/ /g;         # collapse internal whitespace
  return $str;
$$;

-- Returning a composite type
CREATE TYPE name_parts AS (first_name text, last_name text);

CREATE FUNCTION split_name(text) RETURNS name_parts
LANGUAGE plperl AS $$
  my ($full) = @_;
  my ($first, $last) = split(/\s+/, $full, 2);
  return { first_name => $first, last_name => $last };
$$;

-- Set-returning function
CREATE FUNCTION perl_generate_series(integer, integer) RETURNS SETOF integer
LANGUAGE plperl AS $$
  my ($start, $stop) = @_;
  for my $i ($start .. $stop) {
    return_next($i);
  }
  return undef;
$$;

-- Trigger function
CREATE FUNCTION perl_audit_trigger() RETURNS trigger
LANGUAGE plperl AS $$
  $_TD->{new}{modified_at} = localtime();
  return "MODIFY";
$$;

Database access uses spi_exec_query:

CREATE FUNCTION perl_row_count(text) RETURNS integer
LANGUAGE plperl AS $$
  my ($table) = @_;
  my $rv = spi_exec_query("SELECT count(*) AS cnt FROM $table");
  return $rv->{rows}[0]{cnt};
$$;

bool_plperl

Mon, 01 Jan 0001 00:00:00 +0000

plperl : transform between bool and plperl

Overview

ID	Extension	Package	Version	Category	License	Language
3261	bool_plperl	plperl	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Requires	plperl
See Also	plperlu
Siblings	plperl hstore_plperl jsonb_plperl

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION bool_plperl CASCADE; -- requires plperl

Usage

bool_plperl: Transform between bool and PL/Perl

Provides a transform for the bool type for PL/Perl. When loaded, PostgreSQL boolean values are automatically converted to Perl native boolean representations and vice versa, instead of being passed as strings.

CREATE EXTENSION bool_plperl;

CREATE FUNCTION check_flag(val boolean) RETURNS text
LANGUAGE plperl TRANSFORM FOR TYPE boolean AS $$
  # val is a native Perl boolean (1 or undef), not a string
  if ($_[0]) {
    return "flag is set";
  }
  return "flag is not set";
$$;

SELECT check_flag(true);   -- flag is set
SELECT check_flag(false);  -- flag is not set

hstore_plperl

Mon, 01 Jan 0001 00:00:00 +0000

plperl : transform between hstore and plperl

Overview

ID	Extension	Package	Version	Category	License	Language
3262	hstore_plperl	plperl	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
Requires	hstore plperl
See Also	hstore_plpython3u hstore_pllua hstore_plperlu hstore
Siblings	plperl bool_plperl jsonb_plperl

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION hstore_plperl CASCADE; -- requires hstore, plperl

Usage

hstore_plperl: Transform between hstore and PL/Perl

Provides a transform for the hstore type for PL/Perl. When loaded, hstore values are automatically converted to Perl hashes and vice versa.

CREATE EXTENSION hstore_plperl;

CREATE FUNCTION hstore_to_text(val hstore) RETURNS text
LANGUAGE plperl TRANSFORM FOR TYPE hstore AS $$
  # val is now a Perl hash reference
  my $result = '';
  for my $key (sort keys %$val) {
    $result .= "$key => $val->{$key}\n";
  }
  return $result;
$$;

CREATE FUNCTION make_hstore(key text, value text) RETURNS hstore
LANGUAGE plperl TRANSFORM FOR TYPE hstore AS $$
  my ($k, $v) = @_;
  return { $k => $v };
$$;

SELECT hstore_to_text('a=>1, b=>2'::hstore);
SELECT make_hstore('color', 'blue');

jsonb_plperl

Mon, 01 Jan 0001 00:00:00 +0000

plperl : transform between jsonb and plperl

Overview

ID	Extension	Package	Version	Category	License	Language
3263	jsonb_plperl	plperl	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----dt-	No	No	No	Yes	no	yes

Relationships
Requires	plperl
See Also	pgbson pgjq jsonb_plperlu jsonb_plpython3u jsquery pg_jsonschema jsonschema pg_projection
Siblings	plperl bool_plperl hstore_plperl

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION jsonb_plperl CASCADE; -- requires plperl

Usage

jsonb_plperl: Transform between jsonb and PL/Perl

Provides a transform for the jsonb type for PL/Perl. When loaded, jsonb values are automatically converted to Perl data structures (hashes, arrays, scalars) and vice versa.

CREATE EXTENSION jsonb_plperl;

CREATE FUNCTION jsonb_keys_sorted(val jsonb) RETURNS text
LANGUAGE plperl TRANSFORM FOR TYPE jsonb AS $$
  # val is now a Perl hash/array reference
  return join(', ', sort keys %$val);
$$;

CREATE FUNCTION build_jsonb(name text, age integer) RETURNS jsonb
LANGUAGE plperl TRANSFORM FOR TYPE jsonb AS $$
  my ($name, $age) = @_;
  return { name => $name, age => $age };
$$;

SELECT jsonb_keys_sorted('{"b": 2, "a": 1, "c": 3}'::jsonb);
SELECT build_jsonb('Alice', 30);

plperlu

Mon, 01 Jan 0001 00:00:00 +0000

plperlu : PL/PerlU untrusted procedural language

Overview

ID	Extension	Package	Version	Category	License	Language
3270	plperlu	plperlu	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
Schemas	`pg_catalog`
Need By	bool_plperlu hstore_plperlu jsonb_plperlu pg_utl_smtp sparql
See Also	plperl plpython3u plluau pltclu plsh pljava plr plxslt plpgsql
Siblings	bool_plperlu jsonb_plperlu hstore_plperlu

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION plperlu;

Usage

plperlu: PL/Perl untrusted procedural language

PL/Perl Untrusted provides full Perl capabilities including external module loading, filesystem access, and network operations. Only superusers can create functions in this language.

CREATE EXTENSION plperlu;

-- Use external CPAN modules
CREATE FUNCTION fetch_url(text) RETURNS text
LANGUAGE plperlu AS $$
  use LWP::Simple;
  my ($url) = @_;
  return get($url);
$$;

-- Read a file from the server filesystem
CREATE FUNCTION read_server_file(text) RETURNS text
LANGUAGE plperlu AS $$
  my ($path) = @_;
  open my $fh, '<', $path or die "Cannot open $path: $!";
  local $/;
  my $content = <$fh>;
  close $fh;
  return $content;
$$;

-- JSON processing with external module
CREATE FUNCTION parse_json_key(text, text) RETURNS text
LANGUAGE plperlu AS $$
  use JSON;
  my ($json_str, $key) = @_;
  my $data = decode_json($json_str);
  return $data->{$key};
$$;

-- Send email using Net::SMTP
CREATE FUNCTION send_email(text, text, text) RETURNS boolean
LANGUAGE plperlu AS $$
  use Net::SMTP;
  my ($to, $subject, $body) = @_;
  my $smtp = Net::SMTP->new('localhost');
  $smtp->mail('postgres@localhost');
  $smtp->to($to);
  $smtp->data();
  $smtp->datasend("Subject: $subject\n\n$body");
  $smtp->dataend();
  $smtp->quit();
  return 1;
$$;

bool_plperlu

Mon, 01 Jan 0001 00:00:00 +0000

plperlu : transform between bool and plperlu

Overview

ID	Extension	Package	Version	Category	License	Language
3271	bool_plperlu	plperlu	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d--	No	No	No	Yes	no	no

Relationships
Requires	plperlu
See Also	plperl
Siblings	plperlu jsonb_plperlu hstore_plperlu

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION bool_plperlu CASCADE; -- requires plperlu

Usage

bool_plperlu: Transform between bool and PL/Perl untrusted

Provides a transform for the bool type for PL/Perl Untrusted. When loaded, PostgreSQL boolean values are automatically converted to Perl native boolean representations and vice versa, instead of being passed as strings.

CREATE EXTENSION bool_plperlu;

CREATE FUNCTION check_flag_u(val boolean) RETURNS text
LANGUAGE plperlu TRANSFORM FOR TYPE boolean AS $$
  # val is a native Perl boolean (1 or undef), not a string
  if ($_[0]) {
    return "flag is set";
  }
  return "flag is not set";
$$;

SELECT check_flag_u(true);   -- flag is set
SELECT check_flag_u(false);  -- flag is not set

jsonb_plperlu

Mon, 01 Jan 0001 00:00:00 +0000

plperlu : transform between jsonb and plperlu

Overview

ID	Extension	Package	Version	Category	License	Language
3272	jsonb_plperlu	plperlu	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d--	No	No	No	Yes	no	no

Relationships
Requires	plperlu
See Also	pgbson pgjq jsonb_plperl jsonb_plpython3u jsquery pg_jsonschema jsonschema pg_projection
Siblings	plperlu bool_plperlu hstore_plperlu

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION jsonb_plperlu CASCADE; -- requires plperlu

Usage

jsonb_plperlu: Transform between jsonb and PL/Perl untrusted

Provides a transform for the jsonb type for PL/Perl Untrusted. When loaded, jsonb values are automatically converted to Perl data structures (hashes, arrays, scalars) and vice versa.

CREATE EXTENSION jsonb_plperlu;

CREATE FUNCTION process_json_u(val jsonb) RETURNS text
LANGUAGE plperlu TRANSFORM FOR TYPE jsonb AS $$
  use JSON;
  # val is already a Perl data structure, re-encode with sorting
  return encode_json($val);
$$;

CREATE FUNCTION build_jsonb_u(name text, age integer) RETURNS jsonb
LANGUAGE plperlu TRANSFORM FOR TYPE jsonb AS $$
  my ($name, $age) = @_;
  return { name => $name, age => $age };
$$;

SELECT process_json_u('{"b": 2, "a": 1}'::jsonb);
SELECT build_jsonb_u('Alice', 30);

hstore_plperlu

Mon, 01 Jan 0001 00:00:00 +0000

plperlu : transform between hstore and plperlu

Overview

ID	Extension	Package	Version	Category	License	Language
3273	hstore_plperlu	plperlu	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d--	No	No	No	Yes	no	no

Relationships
Requires	hstore plperlu
See Also	hstore_plpython3u hstore_pllua hstore_plperl hstore
Siblings	plperlu bool_plperlu jsonb_plperlu

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION hstore_plperlu CASCADE; -- requires hstore, plperlu

Usage

hstore_plperlu: Transform between hstore and PL/Perl untrusted

Provides a transform for the hstore type for PL/Perl Untrusted. When loaded, hstore values are automatically converted to Perl hashes and vice versa.

CREATE EXTENSION hstore_plperlu;

CREATE FUNCTION hstore_to_json_u(val hstore) RETURNS text
LANGUAGE plperlu TRANSFORM FOR TYPE hstore AS $$
  use JSON;
  # val is now a Perl hash reference
  return encode_json($val);
$$;

CREATE FUNCTION make_hstore_u(key text, value text) RETURNS hstore
LANGUAGE plperlu TRANSFORM FOR TYPE hstore AS $$
  my ($k, $v) = @_;
  return { $k => $v };
$$;

SELECT hstore_to_json_u('a=>1, b=>2'::hstore);
SELECT make_hstore_u('color', 'blue');

plpgsql

Mon, 01 Jan 0001 00:00:00 +0000

plpgsql : PL/pgSQL procedural language

Overview

ID	Extension	Package	Version	Category	License	Language
3280	plpgsql	plpgsql	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Schemas	`pg_catalog`
Need By	bedquilt biscuit cat_tools check_orapg currency data_historization db2fce dbpatch dbstat ddl_historization drop_role_helper dsef event_manager explanation firefly geekspeak generic_plan gogudb grants_manager hello-world hybrid_search index_analyzer istoria italian_codes job_queue json_query json_utils jsonb_schema jx_io keyhippo kilobase kissfft lab-orders launchql-base32 launchql-ext-types launchql-extension-utils launchql-extension-verify launchql-inflection launchql-jwt-claims launchql-stamps launchql-totp livewire medications merge_ips meta_triggers migration monitoring_role mv_rewrite mv_stats myhelper mypg_sharding mysqlcompat newsfeeds nfiesta_gisdata nfiesta_sdesign nfiesta_target_data nonoms norm npm ollama omnidb_plpgsql_debugger partman_to_cstore pase pathman_sharding patients pg-audit-json pg2podg pg4ml pgAutomator pg_abris pg_accumulator pg_audit pg_audit_tools pg_biscuit pg_bleve pg_bm25 pg_cache_tree pg_calcpi pg_catalog_get_defs pg_column_tetris pg_credereum pg_datatype_password pg_dbo_timestamp pg_dbwa pg_dms pg_drop_events pg_dropbuffers pg_dropcache pg_eyes pg_fairmlq pg_fsql pg_gen_uid pg_git pg_graphql_server pg_gsl pg_idm pg_idx_advisor pg_lake_iceberg pg_landmetrics pg_ledger pg_linegazer pg_llm_helper pg_lock_pool pg_message_queue pg_monitoring pg_normalize_email pg_once pg_os pg_osgr pg_pageprep pg_part pg_particulous pg_partman pg_pathman pg_paxos pg_popyramids_datamarts pg_profile pg_prometheus pg_prttn_tools pg_reversi pg_sakila_db pg_semantic_cache pg_sendmail pg_sentence_transformer pg_sessions pg_shardman pg_statviz pg_tileless pg_tms pg_turboquant pg_twkb pg_upless pg_zlog pgaut pgcat pgeyes pgfsm pgh_consistency pgh_hgm pgh_output pgh_output_en_au pgh_output_pt_br pgh_raster pghydro pgmemento pgmock pgnats pgparts pgpm-base32 pgpm-defaults pgpm-faker pgpm-inflection pgpm-jwt-claims pgpm-measurements pgpm-types pgpm-verify pgrao pgrollup pgrouting pgsqlmock pgtap pgtap_fixture pgtelemetry pgvroom plparrot plpgsql_check plpgsql_wrap plrust postgres_ci postpic powa prescriptions qgres quria range_partitioning recall recursively_delete rep_fdw rls_helpers roleman rpg rtiles scheduling session_variables short_ids skitch-extension-defaults skitch-extension-jobs skitch-extension-utils skitch-extension-verify sphinxlink sql_saga supa_queue supabase supabase_auth_apikey sys_syn_dblink tab_tier table_log_pl table_version tablelog telephone test_factory time_for_keys timestampandtz town types unit units us-states uuidv7-sql variant vectors vrprouting wasm webauthn xl_global_views zombodb
See Also	plx plisql plpgsql_check pldbgapi plpgsql_wrap orafce plprofiler plsh pljava pglinter

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION plpgsql;

Usage

plpgsql: PL/pgSQL procedural language

PL/pgSQL is PostgreSQL’s default procedural language. It extends SQL with control structures, variables, cursors, and exception handling.

CREATE EXTENSION plpgsql;  -- installed by default

-- Basic function with variables and control flow
CREATE FUNCTION calculate_discount(price numeric, quantity integer) RETURNS numeric
LANGUAGE plpgsql AS $$
DECLARE
  discount numeric := 0;
BEGIN
  IF quantity >= 100 THEN
    discount := 0.20;
  ELSIF quantity >= 50 THEN
    discount := 0.10;
  ELSIF quantity >= 10 THEN
    discount := 0.05;
  END IF;
  RETURN price * quantity * (1 - discount);
END;
$$;

-- Loop and set-returning function
CREATE FUNCTION fibonacci(n integer) RETURNS SETOF integer
LANGUAGE plpgsql AS $$
DECLARE
  a integer := 0;
  b integer := 1;
  tmp integer;
BEGIN
  FOR i IN 1..n LOOP
    RETURN NEXT a;
    tmp := a + b;
    a := b;
    b := tmp;
  END LOOP;
END;
$$;

SELECT * FROM fibonacci(10);

-- Exception handling
CREATE FUNCTION safe_divide(a numeric, b numeric) RETURNS numeric
LANGUAGE plpgsql AS $$
BEGIN
  RETURN a / b;
EXCEPTION
  WHEN division_by_zero THEN
    RAISE NOTICE 'Division by zero, returning NULL';
    RETURN NULL;
END;
$$;

-- Trigger function
CREATE FUNCTION update_modified_column() RETURNS trigger
LANGUAGE plpgsql AS $$
BEGIN
  NEW.modified_at = now();
  RETURN NEW;
END;
$$;

CREATE TRIGGER set_modified
  BEFORE UPDATE ON my_table
  FOR EACH ROW EXECUTE FUNCTION update_modified_column();

-- Procedure with transaction control (PG 11+)
CREATE PROCEDURE batch_archive(batch_size integer)
LANGUAGE plpgsql AS $$
DECLARE
  rows_moved integer;
BEGIN
  LOOP
    WITH moved AS (
      DELETE FROM orders WHERE status = 'completed'
      RETURNING *
    )
    INSERT INTO orders_archive SELECT * FROM moved;

    GET DIAGNOSTICS rows_moved = ROW_COUNT;
    COMMIT;
    EXIT WHEN rows_moved < batch_size;
  END LOOP;
END;
$$;

plpython3u

Mon, 01 Jan 0001 00:00:00 +0000

plpython3u : PL/Python3U untrusted procedural language

Overview

ID	Extension	Package	Version	Category	License	Language
3290	plpython3u	plpython3u	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
Schemas	`pg_catalog`
Need By	hstore_plpython3u jsonb_plpython3u ltree_plpython3u omni_python pg4ml
See Also	plperl plperlu pllua pltcl plsh pljava plr plxslt plpgsql plv8
Siblings	jsonb_plpython3u ltree_plpython3u hstore_plpython3u

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION plpython3u;

Usage

plpython3u: PL/Python3 untrusted procedural language

PL/Python3U allows writing PostgreSQL functions in Python 3. It is an untrusted language with full access to the Python ecosystem. Only superusers can create functions.

CREATE EXTENSION plpython3u;

-- Simple function
CREATE FUNCTION py_hello(name text) RETURNS text
LANGUAGE plpython3u AS $$
  return f"Hello, {name}!"
$$;

SELECT py_hello('world');

-- Using Python standard library
CREATE FUNCTION py_sha256(data text) RETURNS text
LANGUAGE plpython3u AS $$
  import hashlib
  return hashlib.sha256(data.encode()).hexdigest()
$$;

-- Returning a composite type
CREATE TYPE address AS (street text, city text, zip text);

CREATE FUNCTION parse_address(raw text) RETURNS address
LANGUAGE plpython3u AS $$
  import re
  m = re.match(r'(.+),\s*(.+)\s+(\d{5})', raw)
  if m:
    return (m.group(1), m.group(2), m.group(3))
  return None
$$;

-- Set-returning function
CREATE FUNCTION py_generate_dates(start text, days int) RETURNS SETOF date
LANGUAGE plpython3u AS $$
  from datetime import datetime, timedelta
  d = datetime.strptime(start, '%Y-%m-%d')
  for i in range(days):
    yield (d + timedelta(days=i)).strftime('%Y-%m-%d')
$$;

-- Database access via plpy
CREATE FUNCTION py_row_count(table_name text) RETURNS bigint
LANGUAGE plpython3u AS $$
  result = plpy.execute(f"SELECT count(*) AS cnt FROM {table_name}")
  return result[0]['cnt']
$$;

-- Using external packages (must be installed on the server)
CREATE FUNCTION py_parse_json(url text) RETURNS jsonb
LANGUAGE plpython3u AS $$
  import json, urllib.request
  response = urllib.request.urlopen(url)
  data = json.loads(response.read())
  return json.dumps(data)
$$;

jsonb_plpython3u

Mon, 01 Jan 0001 00:00:00 +0000

plpython3u : transform between jsonb and plpython3u

Overview

ID	Extension	Package	Version	Category	License	Language
3291	jsonb_plpython3u	plpython3u	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d-r	No	No	No	Yes	yes	no

Relationships
Requires	plpython3u
See Also	pgbson pgjq jsonb_plperl jsonb_plperlu jsquery pg_jsonschema jsonschema pg_projection
Siblings	plpython3u ltree_plpython3u hstore_plpython3u

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION jsonb_plpython3u CASCADE; -- requires plpython3u

Usage

jsonb_plpython3u: Transform between jsonb and PL/Python3

Provides a transform for the jsonb type for PL/Python3U. When loaded, jsonb values are automatically converted to Python dicts, lists, and scalars, and vice versa.

CREATE EXTENSION jsonb_plpython3u;

CREATE FUNCTION jsonb_keys_sorted(val jsonb) RETURNS text[]
LANGUAGE plpython3u TRANSFORM FOR TYPE jsonb AS $$
  # val is now a Python dict
  return sorted(val.keys())
$$;

CREATE FUNCTION build_jsonb(name text, age integer) RETURNS jsonb
LANGUAGE plpython3u TRANSFORM FOR TYPE jsonb AS $$
  return {"name": name, "age": age}
$$;

SELECT jsonb_keys_sorted('{"b": 2, "a": 1, "c": 3}'::jsonb);
SELECT build_jsonb('Alice', 30);

ltree_plpython3u

Mon, 01 Jan 0001 00:00:00 +0000

plpython3u : transform between ltree and plpython3u

Overview

ID	Extension	Package	Version	Category	License	Language
3292	ltree_plpython3u	plpython3u	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d-r	No	Yes	No	Yes	yes	no

Relationships
Requires	ltree plpython3u
See Also	ltree
Siblings	plpython3u jsonb_plpython3u hstore_plpython3u

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION ltree_plpython3u CASCADE; -- requires ltree, plpython3u

Usage

ltree_plpython3u: Transform between ltree and PL/Python3

Provides a transform for the ltree type for PL/Python3U. When loaded, ltree values are automatically converted to Python strings and vice versa.

CREATE EXTENSION ltree_plpython3u;

CREATE FUNCTION ltree_depth(val ltree) RETURNS integer
LANGUAGE plpython3u TRANSFORM FOR TYPE ltree AS $$
  # val is now a Python string like "a.b.c"
  return len(val.split('.'))
$$;

CREATE FUNCTION ltree_leaf(val ltree) RETURNS text
LANGUAGE plpython3u TRANSFORM FOR TYPE ltree AS $$
  return val.split('.')[-1]
$$;

SELECT ltree_depth('top.science.astronomy'::ltree);  -- 3
SELECT ltree_leaf('top.science.astronomy'::ltree);    -- astronomy

hstore_plpython3u

Mon, 01 Jan 0001 00:00:00 +0000

plpython3u : transform between hstore and plpython3u

Overview

ID	Extension	Package	Version	Category	License	Language
3293	hstore_plpython3u	plpython3u	`1.0`	LANG	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d-r	No	No	No	Yes	yes	no

Relationships
Requires	hstore plpython3u
See Also	hstore_pllua hstore_plperl hstore_plperlu hstore
Siblings	plpython3u jsonb_plpython3u ltree_plpython3u

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION hstore_plpython3u CASCADE; -- requires hstore, plpython3u

Usage

hstore_plpython3u: Transform between hstore and PL/Python3

Provides a transform for the hstore type for PL/Python3U. When loaded, hstore values are automatically converted to Python dicts and vice versa.

CREATE EXTENSION hstore_plpython3u;

CREATE FUNCTION hstore_to_pairs(val hstore) RETURNS text
LANGUAGE plpython3u TRANSFORM FOR TYPE hstore AS $$
  # val is now a Python dict
  return ', '.join(f'{k}={v}' for k, v in sorted(val.items()))
$$;

CREATE FUNCTION make_hstore(key text, value text) RETURNS hstore
LANGUAGE plpython3u TRANSFORM FOR TYPE hstore AS $$
  return {key: value}
$$;

SELECT hstore_to_pairs('a=>1, b=>2'::hstore);
SELECT make_hstore('color', 'blue');

isn

Mon, 01 Jan 0001 00:00:00 +0000

isn : data types for international product numbering standards

Overview

ID	Extension	Package	Version	Category	License	Language
3930	isn	isn	`1.2`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	country pg_xenophile icu_ext currency asn1oid uri gb18030_2022 pg_cjk_parser

Packages

PG18	PG17	PG16	PG15	PG14
1.2	1.2	1.2	1.2	1.2

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION isn;

Usage

isn: ISBN, ISSN, EAN, UPC product number types

The isn extension provides data types for international product numbering standards with validation and check-digit verification.

CREATE EXTENSION isn;

Data Types

Type	Description
`EAN13`	European Article Numbers (always 13-digit)
`ISBN13`	International Standard Book Numbers (13-digit)
`ISMN13`	International Standard Music Numbers (13-digit)
`ISSN13`	International Standard Serial Numbers (13-digit)
`ISBN`	Book numbers (10-digit short format when possible)
`ISMN`	Music numbers (10-digit short format when possible)
`ISSN`	Serial numbers (10-digit short format when possible)
`UPC`	Universal Product Codes

Examples

SELECT isbn('978-0-393-04002-9');
SELECT isbn13('0901690546');
SELECT issn('1436-4522');

-- Create table with ISBN column
CREATE TABLE books (id isbn);
INSERT INTO books VALUES ('9780393040029');

-- Auto-calculate check digit with ?
INSERT INTO books VALUES ('220500896?');

-- Type casting
SELECT ean13('0901690546'::isbn);

Weak Mode

Enable weak mode to accept invalid check digits (useful for OCR/data cleanup):

SET isn.weak TO true;
INSERT INTO books VALUES ('978-0-11-000533-4');
SET isn.weak TO false;

-- Find invalid entries
SELECT id FROM books WHERE NOT is_valid(id);

-- Fix invalid entries
UPDATE books SET id = make_valid(id) WHERE id = '2-205-00876-X!';

Functions

Function	Description
`is_valid(isn)`	Check if value has valid check digit
`make_valid(isn)`	Clear the invalid-check-digit flag
`isn_weak()`	Return current weak mode status
`isn_weak(boolean)`	Set weak mode

Operators and Indexing

Standard comparison operators (=, <, >, <=, >=, <>) with B-tree and hash index support.

seg

Mon, 01 Jan 0001 00:00:00 +0000

seg : data type for representing line segments or floating-point intervals

Overview

ID	Extension	Package	Version	Category	License	Language
3940	seg	seg	`1.4`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	cube prefix unit ip4r pgmp numeral pg_rational uint uint128 ltree

Packages

PG18	PG17	PG16	PG15	PG14
1.4	1.4	1.4	1.4	1.4

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION seg;

Usage

seg: line segment / floating-point interval data type

The seg extension provides a data type for representing line segments or floating-point intervals, useful for representing measurements with uncertainty.

CREATE EXTENSION seg;

Input Formats

SELECT '5.0'::seg;           -- zero-length interval (point)
SELECT '5.0 .. 7.0'::seg;    -- interval from 5.0 to 7.0
SELECT '5(+-)0.3'::seg;      -- interval 4.7 .. 5.3
SELECT '50 ..'::seg;          -- open interval >= 50
SELECT '.. 0'::seg;           -- open interval <= 0

Certainty indicators (<, >, ~) can be prepended but are treated as comments and ignored by operators.

Operators

Spatial operators (GiST-indexed):

Operator	Description
`<<`	Entirely left of
`>>`	Entirely right of
`&<`	Does not extend right of
`&>`	Does not extend left of
`=`	Equal
`&&`	Overlaps
`@>`	Contains
`<@`	Contained in

Comparison operators (<, <=, >, >=) are available for sorting.

Index Support

CREATE INDEX idx ON measurements USING gist (reading);

Precision

Values are stored as 32-bit floating-point pairs, retaining up to 7 significant digits. Trailing zeroes are preserved.

cube

Mon, 01 Jan 0001 00:00:00 +0000

cube : data type for multidimensional cubes

Overview

ID	Extension	Package	Version	Category	License	Language
3950	cube	cube	`1.5`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	earthdistance pg4ml
See Also	seg prefix unit ip4r pgmp numeral pg_rational uint uint128 ltree

Packages

PG18	PG17	PG16	PG15	PG14
1.5	1.5	1.5	1.5	1.5

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION cube;

Usage

cube: multi-dimensional cube data type

The cube extension provides a data type for representing multi-dimensional cubes (up to 100 dimensions).

CREATE EXTENSION cube;

Input Formats

SELECT '(1,2),(3,4)'::cube;         -- 2D cube
SELECT '(1,2,3)'::cube;              -- 3D point (zero-volume)
SELECT cube(1.0, 2.0);               -- 1D cube from 1 to 2
SELECT cube(ARRAY[1,2], ARRAY[3,4]); -- 2D cube from arrays

Operators

Operator	Description
`&&`	Overlaps
`@>`	Contains
`<@`	Contained in
`->`	Extract n-th coordinate
`<->`	Euclidean distance
`<#>`	Taxicab (L-1) distance
`<=>`	Chebyshev (L-inf) distance

Functions

SELECT cube_dim('(1,2),(3,4)'::cube);                  -- 2
SELECT cube_ll_coord('(1,2),(3,4)'::cube, 1);          -- 1
SELECT cube_ur_coord('(1,2),(3,4)'::cube, 1);          -- 3
SELECT cube_is_point(cube(1,1));                        -- true
SELECT cube_distance('(1,2)'::cube, '(3,4)'::cube);    -- 2.828...
SELECT cube_union('(1,2)'::cube, '(3,4)'::cube);       -- (1,2),(3,4)
SELECT cube_inter('(0,0),(2,2)'::cube, '(1,1),(3,3)'::cube);
SELECT cube_subset(cube('(1,3,5),(6,7,8)'), ARRAY[2]); -- (3),(7)
SELECT cube_enlarge('(1,2),(3,4)'::cube, 0.5, 2);      -- expand by radius

GiST Index with KNN Search

CREATE INDEX idx ON points USING gist (location);

-- Find nearest point to (0.5, 0.5, 0.5)
SELECT * FROM points
ORDER BY location <-> cube(ARRAY[0.5, 0.5, 0.5])
LIMIT 1;

Dimension Handling

When operating on cubes of different dimensions, the lower-dimensional cube is treated as having zeroes for missing coordinates.

ltree

Mon, 01 Jan 0001 00:00:00 +0000

ltree : data type for hierarchical tree-like structures

Overview

ID	Extension	Package	Version	Category	License	Language
3960	ltree	ltree	`1.3`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	fsm_core ltree_plpython3u
See Also	age pg_liquid onesparse graph pgrdf ltree_plpython3u prefix ip4r seg cube

Packages

PG18	PG17	PG16	PG15	PG14
1.3	1.3	1.3	1.3	1.3

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION ltree;

Usage

ltree: hierarchical tree-like label data type

The ltree extension provides data types for hierarchical tree-structured data with extensive search facilities.

CREATE EXTENSION ltree;

Data Types

ltree: Label path (e.g., Top.Science.Astronomy)
lquery: Regex-like pattern for matching ltree values
ltxtquery: Full-text-search-like pattern

Basic Usage

CREATE TABLE categories (path ltree);
INSERT INTO categories VALUES
    ('Top'), ('Top.Science'), ('Top.Science.Astronomy'),
    ('Top.Hobbies'), ('Top.Collections.Pictures');

-- Find descendants
SELECT path FROM categories WHERE path <@ 'Top.Science';

-- Pattern matching
SELECT path FROM categories WHERE path ~ '*.Astronomy.*';

-- Full-text search
SELECT path FROM categories WHERE path @ 'Science & !Pictures';

Operators

Operator	Description
`@>`	Is ancestor of (or equal)
`<@`	Is descendant of (or equal)
`~`	Matches lquery pattern
`?`	Matches any lquery in array
`@`	Matches ltxtquery
`\|\|`	Concatenate paths

lquery Patterns

'*.Science.*'           -- any path containing Science
'Top.*{1,2}.Astronomy'  -- 1-2 labels between Top and Astronomy
'*.astro*'              -- prefix matching
'*.Astro*@'             -- case-insensitive prefix

Functions

SELECT nlevel('Top.Science.Astronomy');                     -- 3
SELECT subltree('Top.Science.Astronomy', 1, 2);            -- Science
SELECT subpath('Top.Science.Astronomy', 1);                 -- Science.Astronomy
SELECT index('a.b.c.d', 'b.c');                             -- 1
SELECT lca('1.2.3', '1.2.3.4.5');                          -- 1.2
SELECT lca(ARRAY['1.2.3'::ltree, '1.2.4'::ltree]);        -- 1.2

Index Support

-- GiST index (supports @>, <@, ~, ?, @)
CREATE INDEX path_gist_idx ON categories USING gist (path);

-- B-tree index (supports <, <=, =, >=, >)
CREATE INDEX path_idx ON categories USING btree (path);

hstore

Mon, 01 Jan 0001 00:00:00 +0000

hstore : data type for storing sets of (key, value) pairs

Overview

ID	Extension	Package	Version	Category	License	Language
3970	hstore	hstore	`1.8`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	flux format hstore_hash_ops hstore_ops hstore_pllua hstore_plluau hstore_plperl hstore_plperlu hstore_plpython2u hstore_plpython3u hstore_plpythonu hstore_plruby json_enhancements_with_hstore mbus meta_triggers numhstore pg_auditor pg_readme pg_rowalesce pg_utility_trigger_functions pg_xenophile sys_syn_dblink
See Also	pgbson collection jsonschema pgjq jsquery pg_jsonschema intarray arraymath pgpdf pglite_fusion

Packages

PG18	PG17	PG16	PG15	PG14
1.8	1.8	1.8	1.8	1.8

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION hstore;

Usage

hstore: key-value pair data type

The hstore extension provides a key/value pair data type for storing semi-structured data in a single column.

CREATE EXTENSION hstore;

Basic Usage

SELECT 'name => Alice, age => 30'::hstore;
SELECT 'name => Alice'::hstore -> 'name';           -- Alice
SELECT 'a => 1, b => 2'::hstore ? 'a';              -- true
SELECT 'a => 1'::hstore || 'b => 2'::hstore;        -- "a"=>"1", "b"=>"2"

Operators

Operator	Description	Example
`->`	Get value by key	`h -> 'key'`
`\|\|`	Concatenate	`h1 \|\| h2`
`?`	Contains key	`h ? 'key'`
`?&`	Contains all keys	`h ?& ARRAY['a','b']`
`?\|`	Contains any key	`h ?\| ARRAY['a','b']`
`@>`	Contains	`h @> 'a=>1'`
`<@`	Contained by	`h <@ 'a=>1, b=>2'`
`-`	Delete key(s)	`h - 'key'` or `h - ARRAY['a','b']`

Subscript Access

SELECT h['name'] FROM mytable;
UPDATE mytable SET h['age'] = '31';

Functions

-- Construction
SELECT hstore('key', 'value');
SELECT hstore(ARRAY['a','b'], ARRAY['1','2']);
SELECT hstore(ROW(1, 'hello'));

-- Extraction
SELECT akeys(h);                    -- text[] of keys
SELECT avals(h);                    -- text[] of values
SELECT skeys(h);                    -- set of keys
SELECT svals(h);                    -- set of values
SELECT each(h);                     -- set of (key, value) records

-- Query
SELECT exist(h, 'key');             -- boolean
SELECT defined(h, 'key');           -- true if non-NULL value

-- Modification
SELECT delete(h, 'key');
SELECT slice(h, ARRAY['a','b']);    -- subset of keys

-- JSON conversion
SELECT hstore_to_json(h);
SELECT hstore_to_jsonb(h);
SELECT hstore_to_json_loose(h);    -- distinguishes numbers/booleans

-- Record conversion
SELECT populate_record(NULL::my_table, h);

Index Support

CREATE INDEX idx ON t USING gin (h);    -- supports @>, ?, ?&, ?|
CREATE INDEX idx ON t USING gist (h);   -- supports @>, ?, ?&, ?|
CREATE INDEX idx ON t USING btree (h);  -- supports =
CREATE INDEX idx ON t USING hash (h);   -- supports =

citext

Mon, 01 Jan 0001 00:00:00 +0000

citext : data type for case-insensitive character strings

Overview

ID	Extension	Package	Version	Category	License	Language
3980	citext	citext	`1.6`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	unaccent icu_ext pg_trgm pg_xenophile gb18030_2022 pg_cjk_parser dict_xsyn pgroonga pg_tokenizer pgpcre

Packages

PG18	PG17	PG16	PG15	PG14
1.6	1.6	1.6	1.6	1.6

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION citext;

Usage

citext: case-insensitive character string type

The citext extension provides a case-insensitive text type that eliminates the need for lower() calls in queries.

CREATE EXTENSION citext;

Basic Usage

CREATE TABLE users (
    nick citext PRIMARY KEY,
    pass text NOT NULL
);

INSERT INTO users VALUES ('Larry', 'secret123');

-- Case-insensitive matching
SELECT * FROM users WHERE nick = 'larry';   -- matches 'Larry'
SELECT * FROM users WHERE nick = 'LARRY';   -- matches 'Larry'

Behavior

citext performs comparisons by internally converting strings to lowercase. The following operations are case-insensitive with citext:

Comparison operators: =, <>, <, >, <=, >=
Pattern matching: LIKE, ILIKE, ~~, ~~*
Regular expressions: ~, ~*, !~, !~*

Case-Insensitive Functions

When arguments are citext, these functions perform case-insensitive matching:

regexp_match(), regexp_matches(), regexp_replace(), regexp_split_to_array(), regexp_split_to_table(), replace(), split_part(), strpos(), translate()

Advantages Over lower()

Eliminates verbose lower() calls in WHERE clauses
Supports case-insensitive PRIMARY KEY and UNIQUE constraints
No need for functional indexes
Transparent case-folding in all operations

Limitations

Case-folding depends on LC_CTYPE at database creation
Slightly less efficient than text (copying and conversion overhead)
Does not support B-tree deduplication
For better Unicode handling, consider nondeterministic collations instead

xml2

Mon, 01 Jan 0001 00:00:00 +0000

xml2 : XPath querying and XSLT

Overview

ID	Extension	Package	Version	Category	License	Language
3990	xml2	xml2	`1.1`	TYPE	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d--	No	No	No	Yes	no	no

Relationships
See Also	plxslt omni_xml pgrdf sparql rdf_fdw pgpdf pgbson hstore pglite_fusion

Packages

PG18	PG17	PG16	PG15	PG14
1.1	1.1	1.1	1.1	1.1

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION xml2;

Usage

xml2: XPath querying and XSLT functionality

The xml2 extension provides XPath querying and XSLT transformation functions for XML documents. Note: this module is deprecated in favor of the core SQL/XML functionality.

CREATE EXTENSION xml2;

XML Validation

SELECT xml_valid('<doc><item>test</item></doc>');  -- true

XPath Query Functions

-- Extract text value
SELECT xpath_string('<doc><name>Alice</name></doc>', '/doc/name');

-- Extract numeric value
SELECT xpath_number('<doc><count>42</count></doc>', '/doc/count');

-- Extract boolean
SELECT xpath_bool('<doc><active>true</active></doc>', '/doc/active');

-- Extract node set with tags
SELECT xpath_nodeset('<doc><a>1</a><a>2</a></doc>', '/doc/a', 'results', 'item');

-- Extract values as comma-separated list
SELECT xpath_list('<doc><a>1</a><a>2</a></doc>', '/doc/a');  -- 1,2

xpath_table Function

Evaluate multiple XPath queries across a set of documents:

SELECT * FROM
  xpath_table('id', 'xml_col', 'documents',
              '/doc/author|/doc/title',
              'true')
  AS t(id int, author text, title text);

Parameters:

Key field name (first output column)
XML document field name
Table/view name
XPath expressions separated by |
WHERE clause (use 'true' for all rows)

XSLT Transformation

-- Apply XSL stylesheet to document
SELECT xslt_process(xml_document, xsl_stylesheet);

-- With parameters
SELECT xslt_process(xml_document, xsl_stylesheet, 'param1=value1,param2=value2');

refint

Mon, 01 Jan 0001 00:00:00 +0000

refint : functions for implementing referential integrity (obsolete)

Overview

ID	Extension	Package	Version	Category	License	Language
4880	refint	refint	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	tcn autoinc insert_username moddatetime table_log

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION refint;

Usage

refint: referential integrity trigger functions

Provides trigger functions for implementing referential integrity checks (largely superseded by built-in foreign key constraints).

CREATE EXTENSION refint;

Trigger Functions

Function	Description
`check_primary_key()`	Check referencing table – ensures referenced row exists
`check_foreign_key()`	Check referenced table – handles cascading actions on delete/update

check_primary_key

Trigger on the referencing table (AFTER INSERT OR UPDATE). Parameters: referencing column name(s), referenced table name, referenced column name(s).

CREATE TRIGGER check_fk
  AFTER INSERT OR UPDATE ON orders
  FOR EACH ROW
  EXECUTE FUNCTION check_primary_key('customer_id', 'customers', 'id');

check_foreign_key

Trigger on the referenced table (AFTER DELETE OR UPDATE). Parameters: number of referencing tables, action (cascade, restrict, or setnull), primary key column(s), then referencing table and column pairs.

CREATE TRIGGER check_ref
  AFTER DELETE OR UPDATE ON customers
  FOR EACH ROW
  EXECUTE FUNCTION check_foreign_key(1, 'cascade', 'id', 'orders', 'customer_id');

autoinc

Mon, 01 Jan 0001 00:00:00 +0000

autoinc : functions for autoincrementing fields

Overview

ID	Extension	Package	Version	Category	License	Language
4881	autoinc	autoinc	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	tcn refint insert_username moddatetime table_log

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION autoinc;

Usage

autoinc: auto-increment trigger function

Provides a trigger function that auto-increments specified columns using sequences.

CREATE EXTENSION autoinc;

Trigger Function

Function	Description
`autoinc()`	Auto-increment columns from sequences on insert (and optionally update)

Parameters are pairs of (column name, sequence name). The value is replaced only if initially zero or NULL.

Examples

CREATE SEQUENCE next_id;
CREATE TABLE ids (id int4, idesc text);

CREATE TRIGGER ids_autoinc
  BEFORE INSERT ON ids
  FOR EACH ROW
  EXECUTE FUNCTION autoinc('id', 'next_id');

INSERT INTO ids (idesc) VALUES ('first');  -- id is auto-assigned from next_id

-- Multiple auto-increment columns
CREATE TRIGGER multi_autoinc
  BEFORE INSERT ON my_table
  FOR EACH ROW
  EXECUTE FUNCTION autoinc('col1', 'seq1', 'col2', 'seq2');

insert_username

Mon, 01 Jan 0001 00:00:00 +0000

insert_username : functions for tracking who changed a table

Overview

ID	Extension	Package	Version	Category	License	Language
4882	insert_username	insert_username	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	tcn refint autoinc moddatetime table_log

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION insert_username;

Usage

insert_username: track who modified a table row

Provides a trigger function that stores the current user’s name into a specified text column.

CREATE EXTENSION insert_username;

Trigger Function

Function	Description
`insert_username()`	Store current username in the specified column

Parameter: name of the text column to store the username.

Examples

CREATE TABLE audit_log (
  id serial PRIMARY KEY,
  data text,
  modified_by text
);

-- Track who inserts
CREATE TRIGGER set_insert_user
  BEFORE INSERT ON audit_log
  FOR EACH ROW
  EXECUTE FUNCTION insert_username('modified_by');

-- Track who updates
CREATE TRIGGER set_update_user
  BEFORE UPDATE ON audit_log
  FOR EACH ROW
  EXECUTE FUNCTION insert_username('modified_by');

INSERT INTO audit_log (data) VALUES ('test');
SELECT modified_by FROM audit_log;  -- returns current user

moddatetime

Mon, 01 Jan 0001 00:00:00 +0000

moddatetime : functions for tracking last modification time

Overview

ID	Extension	Package	Version	Category	License	Language
4883	moddatetime	moddatetime	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	tcn refint autoinc insert_username table_log

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION moddatetime;

Usage

moddatetime: track modification timestamp

Provides a trigger function that stores the current timestamp when a row is modified.

CREATE EXTENSION moddatetime;

Trigger Function

Function	Description
`moddatetime()`	Store current timestamp in the specified column on UPDATE

Parameter: name of the timestamp or timestamp with time zone column to update.

Examples

CREATE TABLE documents (
  id serial PRIMARY KEY,
  content text,
  modified_at timestamp with time zone
);

CREATE TRIGGER set_modified
  BEFORE UPDATE ON documents
  FOR EACH ROW
  EXECUTE FUNCTION moddatetime('modified_at');

UPDATE documents SET content = 'new content' WHERE id = 1;
-- modified_at is automatically set to current timestamp

tsm_system_time

Mon, 01 Jan 0001 00:00:00 +0000

tsm_system_time : TABLESAMPLE method which accepts time in milliseconds as a limit

Overview

ID	Extension	Package	Version	Category	License	Language
4890	tsm_system_time	tsm_system_time	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	faker dbt2 random tsm_system_rows pg_mockable pgsqlmock pgtap omni_test

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION tsm_system_time;

Usage

tsm_system_time: time-based TABLESAMPLE method

Provides the SYSTEM_TIME table sampling method that returns as many rows as can be read within a specified time limit.

CREATE EXTENSION tsm_system_time;

TABLESAMPLE Method

SYSTEM_TIME(milliseconds float) – maximum time to spend reading the table.

Examples

-- Sample rows readable within 1 second
SELECT * FROM my_table TABLESAMPLE SYSTEM_TIME(1000);

-- Sample from a large table with a 500ms budget
SELECT count(*) FROM large_table TABLESAMPLE SYSTEM_TIME(500);

Performs block-level sampling (not row-level). If the entire table can be read within the time limit, all rows are returned. Does not support REPEATABLE.

dict_xsyn

Mon, 01 Jan 0001 00:00:00 +0000

dict_xsyn : text search dictionary template for extended synonym processing

Overview

ID	Extension	Package	Version	Category	License	Language
4900	dict_xsyn	dict_xsyn	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	dict_int unaccent pg_tokenizer pgroonga pg_kazsearch pg_jieba pg_cjk_parser zhparser

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION dict_xsyn;

Usage

dict_xsyn: extended synonym dictionary for text search

Provides an extended synonym dictionary template for text search, replacing words with groups of synonyms.

CREATE EXTENSION dict_xsyn;

Configuration Parameters

Parameter	Description	Default
`matchorig`	Accept original word	`true`
`matchsynonyms`	Accept synonyms as input	`false`
`keeporig`	Include original in output	`true`
`keepsynonyms`	Include synonyms in output	`true`
`rules`	Base name of synonym file in `$SHAREDIR/tsearch_data/` (`.rules` extension)	–

Rules File Format

word syn1 syn2 syn3

Lines starting with # are comments.

Examples

-- Configure the dictionary
ALTER TEXT SEARCH DICTIONARY xsyn (RULES='my_rules', KEEPORIG=true);

-- Test the dictionary
SELECT ts_lexize('xsyn', 'word');
-- {word,syn1,syn2,syn3}

-- Match synonyms as input too
ALTER TEXT SEARCH DICTIONARY xsyn (RULES='my_rules', MATCHSYNONYMS=true);
SELECT ts_lexize('xsyn', 'syn1');
-- {syn1,syn2,syn3}

-- Use in a text search configuration
ALTER TEXT SEARCH CONFIGURATION english
  ALTER MAPPING FOR word, asciiword WITH xsyn, english_stem;

tsm_system_rows

Mon, 01 Jan 0001 00:00:00 +0000

tsm_system_rows : TABLESAMPLE method which accepts number of rows as a limit

Overview

ID	Extension	Package	Version	Category	License	Language
4910	tsm_system_rows	tsm_system_rows	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	documentdb
See Also	faker dbt2 random tsm_system_time pg_mockable pgsqlmock pgtap omni_test

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION tsm_system_rows;

Usage

tsm_system_rows: row-count-based TABLESAMPLE method

Provides the SYSTEM_ROWS table sampling method that returns exactly the specified number of rows.

CREATE EXTENSION tsm_system_rows;

TABLESAMPLE Method

SYSTEM_ROWS(count int) – maximum number of rows to return.

Examples

-- Sample exactly 100 rows
SELECT * FROM my_table TABLESAMPLE SYSTEM_ROWS(100);

-- Quick peek at 10 rows from a large table
SELECT * FROM large_table TABLESAMPLE SYSTEM_ROWS(10);

Performs block-level sampling (may exhibit clustering effects with small samples). Returns all rows if the table has fewer rows than requested. Does not support REPEATABLE.

tcn

Mon, 01 Jan 0001 00:00:00 +0000

tcn : Triggered change notifications

Overview

ID	Extension	Package	Version	Category	License	Language
4920	tcn	tcn	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	pgmq pgmb ulak redis pgq pgmqtt pg_durable redis_fdw fsm_core refint

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION tcn;

Usage

tcn: triggered change notifications via LISTEN/NOTIFY

Provides a trigger function that sends NOTIFY events with information about changed rows, enabling asynchronous change tracking.

CREATE EXTENSION tcn;

Trigger Function

Function	Description
`triggered_change_notification()`	Send NOTIFY on row changes with primary key info

Optional parameter: custom channel name (defaults to tcn).

Notification Payload Format

"table_name",operation,"column"='value',"column"='value'

Operations: I (INSERT), U (UPDATE), D (DELETE).

Examples

CREATE TABLE tcndata (
  a int NOT NULL,
  b date NOT NULL,
  c text,
  PRIMARY KEY (a, b)
);

-- Attach the trigger
CREATE TRIGGER tcndata_tcn
  AFTER INSERT OR UPDATE OR DELETE ON tcndata
  FOR EACH ROW
  EXECUTE FUNCTION triggered_change_notification();

-- Listen for notifications
LISTEN tcn;

-- Changes trigger notifications:
INSERT INTO tcndata VALUES (1, '2024-01-01', 'test');
-- Notification: "tcndata",I,"a"='1',"b"='2024-01-01'

UPDATE tcndata SET c = 'updated' WHERE a = 1;
-- Notification: "tcndata",U,"a"='1',"b"='2024-01-01'

DELETE FROM tcndata WHERE a = 1;
-- Notification: "tcndata",D,"a"='1',"b"='2024-01-01'

-- Use a custom channel name
CREATE TRIGGER my_trigger
  AFTER INSERT OR UPDATE OR DELETE ON my_table
  FOR EACH ROW
  EXECUTE FUNCTION triggered_change_notification('my_channel');

uuid-ossp

Mon, 01 Jan 0001 00:00:00 +0000

uuid-ossp : generate universally unique identifiers (UUIDs)

Overview

ID	Extension	Package	Version	Category	License	Language
4930	uuid-ossp	uuid-ossp	`1.1`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	babelfishpg_tsql bundle datalink launchql-extension-verify launchql-inflection launchql-jwt-claims npm pg_abris pg_dms provsql ruid skitch-extension-jobs skitch-extension-verify supa_audit types
See Also	pg_uuidv7 sequential_uuids pg_idkit pgx_ulid snowflake pg_uuid_v8 typeid

Packages

PG18	PG17	PG16	PG15	PG14
1.1	1.1	1.1	1.1	1.1

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION uuid-ossp;

Usage

uuid-ossp: UUID generation functions

Provides functions to generate UUIDs using several standard algorithms. Note: for simple random UUIDs, consider using the built-in gen_random_uuid() instead.

CREATE EXTENSION "uuid-ossp";

UUID Generation Functions

Function	Description
`uuid_generate_v1()`	Version 1: MAC address + timestamp
`uuid_generate_v1mc()`	Version 1 with random multicast MAC
`uuid_generate_v3(namespace uuid, name text)`	Version 3: MD5 hash of namespace + name
`uuid_generate_v4()`	Version 4: fully random
`uuid_generate_v5(namespace uuid, name text)`	Version 5: SHA-1 hash of namespace + name (preferred over v3)

Namespace Constants

Function	Description
`uuid_nil()`	Nil UUID (all zeros)
`uuid_ns_dns()`	DNS namespace
`uuid_ns_url()`	URL namespace
`uuid_ns_oid()`	ISO OID namespace
`uuid_ns_x500()`	X.500 DN namespace

Examples

-- Random UUID (v4)
SELECT uuid_generate_v4();

-- Timestamp-based UUID (v1)
SELECT uuid_generate_v1();

-- Deterministic UUID from name (v5, preferred over v3)
SELECT uuid_generate_v5(uuid_ns_url(), 'http://www.postgresql.org');

-- Use as default primary key
CREATE TABLE items (
  id uuid DEFAULT uuid_generate_v4() PRIMARY KEY,
  name text
);

btree_gist

Mon, 01 Jan 0001 00:00:00 +0000

btree_gist : support for indexing common datatypes in GiST

Overview

ID	Extension	Package	Version	Category	License	Language
4940	btree_gist	btree_gist	`1.7`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
Need By	emaj omni_auth periods pg_lake_table pgautofailover powa
See Also	ora_btree_gist btree_gin parray_gin intarray rum bloom db2fce orafce

Packages

PG18	PG17	PG16	PG15	PG14
1.7	1.7	1.7	1.7	1.7

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION btree_gist;

Usage

btree_gist: B-tree equivalent GiST operator classes

Provides GiST index operator classes for data types that normally only support B-tree indexing. Enables exclusion constraints combining equality with range operators.

CREATE EXTENSION btree_gist;

Supported Data Types

int2, int4, int8, float4, float8, numeric, timestamp with time zone, timestamp without time zone, time with time zone, time without time zone, date, interval, oid, money, char, varchar, text, bytea, bit, varbit, macaddr, macaddr8, inet, cidr, uuid, bool, and all enum types.

Distance Operator

The <-> operator is provided for nearest-neighbor searches on numeric and temporal types.

Examples

-- GiST index on integer column
CREATE INDEX idx ON test USING GIST (a);
SELECT * FROM test WHERE a < 10;

-- Nearest-neighbor search
SELECT *, a <-> 42 AS dist FROM test ORDER BY a <-> 42 LIMIT 10;

-- Exclusion constraint: each cage can only contain one type of animal
CREATE TABLE zoo (
  cage   integer,
  animal text,
  EXCLUDE USING GIST (cage WITH =, animal WITH <>)
);

INSERT INTO zoo VALUES (1, 'lion');    -- OK
INSERT INTO zoo VALUES (1, 'tiger');   -- ERROR: conflicting key value
INSERT INTO zoo VALUES (2, 'tiger');   -- OK

-- Exclusion constraint for non-overlapping time ranges per room
CREATE TABLE reservations (
  room int,
  during tsrange,
  EXCLUDE USING GIST (room WITH =, during WITH &&)
);

btree_gin

Mon, 01 Jan 0001 00:00:00 +0000

btree_gin : support for indexing common datatypes in GIN

Overview

ID	Extension	Package	Version	Category	License	Language
4950	btree_gin	btree_gin	`1.3`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	ora_btree_gin btree_gist parray_gin intarray rum bloom db2fce orafce

Packages

PG18	PG17	PG16	PG15	PG14
1.3	1.3	1.3	1.3	1.3

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION btree_gin;

Usage

btree_gin: B-tree equivalent GIN operator classes

Provides GIN index operator classes for data types that normally only support B-tree indexing. Useful for multicolumn GIN indexes that combine GIN-indexable and B-tree-indexable columns.

CREATE EXTENSION btree_gin;

Supported Data Types

int2, int4, int8, float4, float8, numeric, timestamp with time zone, timestamp without time zone, time with time zone, time without time zone, date, interval, oid, money, char, varchar, text, bytea, macaddr, macaddr8, inet, cidr, uuid, bit, varbit, bool, name, bpchar, and all enum types.

Examples

-- GIN index on an integer column
CREATE INDEX idx ON test USING GIN (a);
SELECT * FROM test WHERE a < 10;

-- Multicolumn GIN index combining full-text search with a scalar filter
CREATE INDEX idx ON articles USING GIN (body_tsvector, category);
SELECT * FROM articles
WHERE body_tsvector @@ to_tsquery('PostgreSQL')
  AND category = 'tech';

Note: btree_gin does not outperform standard B-tree indexes for single-column queries. Its main benefit is combining scalar columns with GIN-native columns (like tsvector or arrays) in a single multicolumn index.

intarray

Mon, 01 Jan 0001 00:00:00 +0000

intarray : functions, operators, and index support for 1-D arrays of integers

Overview

ID	Extension	Package	Version	Category	License	Language
4960	intarray	intarray	`1.5`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----dt-	No	No	No	Yes	no	yes

Relationships
See Also	arraymath floatvec aggs_for_vecs aggs_for_arrays collection hstore pgbson parray_gin btree_gin btree_gist

Packages

PG18	PG17	PG16	PG15	PG14
1.5	1.5	1.5	1.5	1.5

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION intarray;

Usage

intarray: integer array functions and operators with index support

Provides functions and operators for manipulating null-free integer arrays, with GiST and GIN index support for fast array searches.

CREATE EXTENSION intarray;

Functions

Function	Description	Example
`icount(int[])`	Number of elements	`icount('{1,2,3}')` – 3
`sort(int[], dir)`	Sort array (`'asc'` or `'desc'`)	`sort('{3,1,2}','asc')` – `{1,2,3}`
`sort_asc(int[])`	Sort ascending	`sort_asc('{3,1,2}')` – `{1,2,3}`
`sort_desc(int[])`	Sort descending	`sort_desc('{3,1,2}')` – `{3,2,1}`
`uniq(int[])`	Remove adjacent duplicates	`uniq(sort('{1,2,3,2,1}'))` – `{1,2,3}`
`idx(int[], item)`	Index of first match	`idx('{11,22,33}', 22)` – 2
`subarray(int[], start, len)`	Extract sub-array	`subarray('{1,2,3,4}', 2, 2)` – `{2,3}`
`intset(int)`	Make single-element array	`intset(42)` – `{42}`

Operators

Operator	Description
`&&`	Arrays overlap (have common elements)
`@>`	Left array contains right
`<@`	Left array is contained in right
`#`	Number of elements
`+`	Array concatenation / append element
`-`	Remove elements
`\|`	Union of arrays
`&`	Intersection of arrays
`@@`	Array matches a query expression
`~~`	Query expression matches array

Index Support

-- GiST index for array containment/overlap queries
CREATE INDEX idx ON messages USING GIST (tags gist__intbig_ops);

-- GIN index (alternative)
CREATE INDEX idx ON messages USING GIN (tags gin__int_ops);

-- Query with index support
SELECT * FROM messages WHERE tags && '{1,2}';     -- overlap
SELECT * FROM messages WHERE tags @> '{1,2}';     -- contains
SELECT * FROM messages WHERE tags @@ '1&(2|3)';  -- query expression

intagg

Mon, 01 Jan 0001 00:00:00 +0000

intagg : integer aggregator and enumerator (obsolete)

Overview

ID	Extension	Package	Version	Category	License	Language
4970	intagg	intagg	`1.1`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
----d--	No	No	No	Yes	no	no

Relationships
See Also	intarray count_distinct arraymath

Packages

PG18	PG17	PG16	PG15	PG14
1.1	1.1	1.1	1.1	1.1

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION intagg;

Usage

intagg: integer aggregator and enumerator (obsolete)

Provides an integer aggregator and enumerator. These are now wrappers around the built-in array_agg() and unnest() functions.

CREATE EXTENSION intagg;

Functions

Function	Description
`int_array_aggregate(integer)`	Aggregate integers into an array (wrapper for `array_agg()`)
`int_array_enum(integer[])`	Expand array into rows (wrapper for `unnest()`)

Examples

-- Aggregate integers into an array
SELECT id_left, int_array_aggregate(id_right) AS rights
FROM many_to_many
GROUP BY id_left;

-- Expand an integer array into rows
SELECT int_array_enum(ARRAY[1, 2, 3, 4]);
-- Returns: 1, 2, 3, 4 (as separate rows)

Note: This module is obsolete. Use the built-in array_agg() and unnest() functions instead for new code.

dict_int

Mon, 01 Jan 0001 00:00:00 +0000

dict_int : text search dictionary template for integers

Overview

ID	Extension	Package	Version	Category	License	Language
4980	dict_int	dict_int	`1.0`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	dict_xsyn unaccent pg_tokenizer pgroonga pg_kazsearch pg_jieba pg_cjk_parser zhparser

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION dict_int;

Usage

dict_int: integer text search dictionary

Controls the indexing of integers in full-text search to prevent excessive unique-word growth that degrades search performance.

CREATE EXTENSION dict_int;

Configuration Parameters

Parameter	Description	Default
`maxlen`	Maximum number of digits allowed	6
`rejectlong`	If `true`, reject overlength integers (stop word). If `false`, truncate.	`false`
`absval`	If `true`, strip leading `+`/`-` before applying maxlen	`false`

Examples

-- Test the default dictionary
SELECT ts_lexize('intdict', '12345678');
-- {123456}  (truncated to 6 digits by default)

-- Configure to reject long integers
ALTER TEXT SEARCH DICTIONARY intdict (MAXLEN = 4, REJECTLONG = true);
SELECT ts_lexize('intdict', '12345');
-- {}  (rejected as stop word)

SELECT ts_lexize('intdict', '1234');
-- {1234}  (accepted)

-- Apply to a text search configuration
ALTER TEXT SEARCH CONFIGURATION english
  ALTER MAPPING FOR int, uint WITH intdict;

unaccent

Mon, 01 Jan 0001 00:00:00 +0000

unaccent : text search dictionary that removes accents

Overview

ID	Extension	Package	Version	Category	License	Language
4990	unaccent	unaccent	`1.1`	FUNC	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	dict_xsyn dict_int pg_tokenizer pgroonga icu_ext pg_xenophile citext pg_trgm

Packages

PG18	PG17	PG16	PG15	PG14
1.1	1.1	1.1	1.1	1.1

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION unaccent;

Usage

unaccent: text search dictionary for accent removal

Removes accents (diacritic signs) from text. Can be used as a text search dictionary or as a standalone function.

CREATE EXTENSION unaccent;

Functions

Function	Description
`unaccent(text)`	Remove accents using the default dictionary
`unaccent(dictionary regdictionary, text)`	Remove accents using a specific dictionary

Text Search Usage

-- Test the dictionary
SELECT ts_lexize('unaccent', 'Hôtel');
-- {Hotel}

-- Create an accent-insensitive French text search config
CREATE TEXT SEARCH CONFIGURATION fr (COPY = french);
ALTER TEXT SEARCH CONFIGURATION fr
  ALTER MAPPING FOR hword, hword_part, word
  WITH unaccent, french_stem;

SELECT to_tsvector('fr', 'Hôtels de la Mer');
-- 'hotel':1 'mer':4

-- Accent-insensitive search
SELECT to_tsvector('fr', 'Hôtel de la Mer') @@ to_tsquery('fr', 'Hotels');
-- true

Standalone Function Usage

SELECT unaccent('Hôtel');           -- Hotel
SELECT unaccent('café');            -- cafe
SELECT unaccent('résumé');          -- resume
SELECT unaccent('niño');            -- nino

pg_prewarm

Mon, 01 Jan 0001 00:00:00 +0000

pg_prewarm : prewarm relation data

Overview

ID	Extension	Package	Version	Category	License	Language
5890	pg_prewarm	pg_prewarm	`1.2`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--sLd--	No	Yes	Yes	Yes	no	no

Relationships
Need By	pgcozy
See Also	pg_buffercache pgfincore pgcozy pg_cooldown

Packages

PG18	PG17	PG16	PG15	PG14
1.2	1.2	1.2	1.2	1.2

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Config this extension to shared_preload_libraries:

shared_preload_libraries = 'pg_prewarm';

Create this extension with:

CREATE EXTENSION pg_prewarm;

Usage

pg_prewarm: prewarm relation data

The pg_prewarm extension provides functions to load relation data into the operating system buffer cache or PostgreSQL buffer cache, reducing I/O latency for subsequent queries.

Prewarm a Relation

-- Load entire table into PostgreSQL buffer cache (default mode)
SELECT pg_prewarm('my_table');

-- Load with specific mode
SELECT pg_prewarm('my_table', 'prefetch');  -- async OS prefetch
SELECT pg_prewarm('my_table', 'read');      -- sync read into OS cache
SELECT pg_prewarm('my_table', 'buffer');    -- load into PG buffer cache

-- Load specific block range
SELECT pg_prewarm('my_table', 'buffer', 'main', 0, 999);

-- Prewarm an index
SELECT pg_prewarm('my_table_pkey');

Function Signature

pg_prewarm(regclass,
           mode text DEFAULT 'buffer',
           fork text DEFAULT 'main',
           first_block int8 DEFAULT NULL,
           last_block int8 DEFAULT NULL
) RETURNS int8

Returns the number of blocks prewarmed.

Parameter	Description
`mode`	`prefetch` (async OS), `read` (sync OS), or `buffer` (PG cache)
`fork`	Relation fork to prewarm (e.g., `main`, `fsm`, `vm`)
`first_block`	First block number (default: 0)
`last_block`	Last block number (default: last block of relation)

Autoprewarm

When loaded via shared_preload_libraries, autoprewarm periodically saves the list of buffers in the shared buffer cache to disk and restores them on restart.

-- Manually launch autoprewarm worker
SELECT autoprewarm_start_worker();

-- Force immediate dump of buffer state
SELECT autoprewarm_dump_now();  -- returns number of records written

GUC Parameters

Parameter	Default	Description
`pg_prewarm.autoprewarm`	`true`	Enable autoprewarm background worker
`pg_prewarm.autoprewarm_interval`	`300s`	Interval between `autoprewarm.blocks` file updates (0 = dump only at shutdown)

Buffer state is saved to autoprewarm.blocks in the data directory. After restart, two background workers reload the saved buffers.

lo

Mon, 01 Jan 0001 00:00:00 +0000

lo : Large Object maintenance

Overview

ID	Extension	Package	Version	Category	License	Language
5930	lo	lo	`1.1`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-dt-	No	Yes	No	Yes	no	yes

Relationships
See Also	lolor pg_surgery pg_repack pglogical spock

Packages

PG18	PG17	PG16	PG15	PG14
1.1	1.1	1.1	1.1	1.1

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION lo;

Usage

lo: Large Object maintenance

The lo extension provides a data type and trigger function for managing PostgreSQL Large Objects, preventing orphaned objects when references are updated or deleted.

Data Type

The lo type is a domain over oid, used to identify columns that hold Large Object references. This is especially useful for ODBC driver compatibility.

CREATE TABLE image (
    title  text,
    raster lo      -- large object reference column
);

Trigger Function

The lo_manage() trigger automatically calls lo_unlink() to delete the associated Large Object when a row is updated or deleted:

CREATE TRIGGER t_raster
    BEFORE UPDATE OR DELETE ON image
    FOR EACH ROW EXECUTE FUNCTION lo_manage(raster);

For multiple lo columns, create a separate trigger for each:

CREATE TABLE gallery (
    title     text,
    thumbnail lo,
    fullsize  lo
);

CREATE TRIGGER t_thumbnail
    BEFORE UPDATE OR DELETE ON gallery
    FOR EACH ROW EXECUTE FUNCTION lo_manage(thumbnail);

CREATE TRIGGER t_fullsize
    BEFORE UPDATE OR DELETE ON gallery
    FOR EACH ROW EXECUTE FUNCTION lo_manage(fullsize);

To restrict the trigger to column updates only:

CREATE TRIGGER t_raster
    BEFORE UPDATE OF raster OR DELETE ON image
    FOR EACH ROW EXECUTE FUNCTION lo_manage(raster);

Limitations

DROP TABLE and TRUNCATE do not fire row-level triggers, so Large Objects will be orphaned. Run DELETE FROM table before dropping.
The trigger assumes each Large Object is referenced by only one column/row.
Use the vacuumlo utility to clean up any orphaned Large Objects.

The extension is trusted and can be installed by non-superusers with CREATE privilege.

basic_archive

Mon, 01 Jan 0001 00:00:00 +0000

basic_archive : an example of an archive module

Overview

ID	Extension	Package	Version	Category	License	Language
5940	basic_archive	basic_archive	`-`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s----	No	Yes	No	No	no	no

Relationships
See Also	aws_s3 basebackup_to_shell external_file file_fdw repmgr pg_parquet pg_bulkload

Packages

PG18	PG17	PG16	PG15	PG14
-	-	-	-	N/A

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

This extension does not need CREATE EXTENSION DDL command

Usage

basic_archive: an example of an archive module

The basic_archive module is a WAL archive module that copies completed WAL segment files to a specified directory. It serves as a reference implementation for custom archive modules.

Configuration

Add to postgresql.conf:

archive_mode = 'on'
archive_library = 'basic_archive'
basic_archive.archive_directory = '/path/to/archive/directory'

Parameters

Parameter	Type	Description
`basic_archive.archive_directory`	string	Directory to copy WAL files to (must already exist)

If archive_mode is enabled but basic_archive.archive_directory is empty (default), the server will accumulate WAL files until a directory path is configured.

Notes

The target directory must be created before use; the module will not create it
After a server crash, temporary files with the archtemp prefix may be left in the archive directory and should be deleted before restarting
These temporary files can also be safely removed while the server is running, provided they are not related to an ongoing archive operation
This module is primarily intended as a simple example and starting point for developing custom archive modules

basebackup_to_shell

Mon, 01 Jan 0001 00:00:00 +0000

basebackup_to_shell : adds a custom basebackup target called shell

Overview

ID	Extension	Package	Version	Category	License	Language
5950	basebackup_to_shell	basebackup_to_shell	`-`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s----	No	Yes	No	No	no	no

Relationships
See Also	basic_archive aws_s3 external_file file_fdw repmgr pg_parquet pg_bulkload

Packages

PG18	PG17	PG16	PG15	PG14
-	-	-	-	N/A

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

This extension does not need CREATE EXTENSION DDL command

Usage

basebackup_to_shell: adds a custom basebackup target called shell

The basebackup_to_shell module adds a custom shell backup target to pg_basebackup, allowing administrators to pipe backup archives through arbitrary shell commands.

Configuration

Add to postgresql.conf:

shared_preload_libraries = 'basebackup_to_shell'

# Command to execute for each archive; receives archive data on stdin
basebackup_to_shell.command = 'gzip > /backup/%f.gz'

# Optional: restrict usage to a specific role
basebackup_to_shell.required_role = 'backup_admin'

Parameters

Parameter	Type	Description
`basebackup_to_shell.command`	string	Shell command to execute; receives archive on stdin
`basebackup_to_shell.required_role`	string	Role required to use the shell target (empty = any replication user)

Command Placeholders

Placeholder	Replaced With
`%f`	Archive filename (e.g., `base.tar`)
`%d`	User-provided target detail string
`%%`	Literal `%`

Examples

# Compress backups to a local directory
# postgresql.conf: basebackup_to_shell.command = 'gzip > /backup/%f.gz'
pg_basebackup --target=shell

# Upload to S3 with a detail parameter
# postgresql.conf: basebackup_to_shell.command = 'aws s3 cp - s3://bucket/%d/%f'
pg_basebackup --target=shell:myprefix

# Custom processing pipeline
# postgresql.conf: basebackup_to_shell.command = 'zstd | ssh backup-host "cat > /backup/%f.zst"'
pg_basebackup --target=shell

The %d placeholder requires a target detail to be provided via --target=shell:DETAIL. If %d is not in the command, providing a detail is prohibited. Detail strings may only contain alphanumeric characters.

old_snapshot

Mon, 01 Jan 0001 00:00:00 +0000

old_snapshot : utilities in support of old_snapshot_threshold

Overview

ID	Extension	Package	Version	Category	License	Language
5960	old_snapshot	old_snapshot	`1.0`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	pg_dirtyread pg_visibility temporal_tables pgstattuple

Packages

PG18	PG17	PG16	PG15	PG14
N/A	N/A	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION old_snapshot;

Usage

old_snapshot: utilities in support of old_snapshot_threshold

The old_snapshot extension provides inspection functions for the server state related to the old_snapshot_threshold configuration parameter.

Note: this chapter is removed from PostgreSQL current docs in newer releases; use versioned docs when needed.

Function

-- View the timestamp-to-XID mapping table
SELECT * FROM pg_old_snapshot_time_mapping();

Function Signature

pg_old_snapshot_time_mapping(
    array_offset OUT int4,
    end_timestamp OUT timestamptz,
    newest_xmin OUT xid
) RETURNS SETOF record

Output Columns

Column	Type	Description
`array_offset`	int4	Index position in the mapping array
`end_timestamp`	timestamptz	End of the corresponding one-minute interval
`newest_xmin`	xid	Newest xmin of any snapshot taken during that minute

Context

PostgreSQL’s old_snapshot_threshold parameter controls how long a snapshot can remain valid. The server maintains an internal mapping of timestamps to transaction IDs to implement this feature. This extension exposes that mapping for inspection and debugging.

-- Check the old_snapshot_threshold setting
SHOW old_snapshot_threshold;

-- Inspect the current mapping entries
SELECT array_offset, end_timestamp, newest_xmin
FROM pg_old_snapshot_time_mapping()
ORDER BY array_offset;

adminpack

Mon, 01 Jan 0001 00:00:00 +0000

adminpack : administrative functions for PostgreSQL

Overview

ID	Extension	Package	Version	Category	License	Language
5970	adminpack	adminpack	`2.1`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	pg_checksums amcheck pg_repack pg_catcheck pg_surgery pg_orphaned pg_pathcheck pg_isok pageinspect pg_squeeze

Packages

PG18	PG17	PG16	PG15	PG14
N/A	N/A	2.1	2.1	2.1

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION adminpack;

Usage

adminpack: administrative functions for PostgreSQL

The adminpack extension provides server-side file management and log access functions, primarily used by pgAdmin and other administration tools. All functions are restricted to superusers by default.

File Operations

-- Write text to a file (append=false: file must not exist; append=true: append)
SELECT pg_file_write('/tmp/test.txt', 'Hello World', false);   -- returns bytes written

-- Append to existing file
SELECT pg_file_write('/tmp/test.txt', E'\nMore data', true);

-- Sync file to disk
SELECT pg_file_sync('/tmp/test.txt');

-- Rename a file
SELECT pg_file_rename('/tmp/old.txt', '/tmp/new.txt');

-- Rename with archiving (moves existing newname to archive first)
SELECT pg_file_rename('/tmp/old.txt', '/tmp/new.txt', '/tmp/archive.txt');

-- Delete a file
SELECT pg_file_unlink('/tmp/test.txt');   -- returns true on success

Log File Access

-- List server log files (requires default log_filename format)
SELECT * FROM pg_logdir_ls();

Returns timestamps and paths of log files from the log_directory.

Function Reference

Function	Returns	Description
`pg_file_write(filename, data, append)`	bigint	Write text to file; returns bytes written
`pg_file_sync(filename)`	void	Flush file or directory to disk
`pg_file_rename(old, new [, archive])`	boolean	Rename file, optionally archiving existing target
`pg_file_unlink(filename)`	boolean	Delete a file
`pg_logdir_ls()`	setof record	List log files with timestamps

Access Control

All functions default to superuser-only access
Permissions can be granted via GRANT to non-superusers
File access is restricted to the database cluster directory unless the user has pg_read_server_files or pg_write_server_files roles

Note: adminpack was removed in PostgreSQL 17. For PostgreSQL 16 and earlier only.

amcheck

Mon, 01 Jan 0001 00:00:00 +0000

amcheck : functions for verifying relation integrity

Overview

ID	Extension	Package	Version	Category	License	Language
5980	amcheck	amcheck	`1.4`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	pg_checksums pg_catcheck pg_orphaned pg_pathcheck pg_isok pg_surgery pageinspect adminpack pg_repack

Packages

PG18	PG17	PG16	PG15	PG14
1.4	1.4	1.4	1.4	1.4

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION amcheck;

Usage

amcheck: functions for verifying relation integrity

The amcheck extension provides functions to verify the logical consistency of B-tree indexes, GIN indexes, and heap (table) data, detecting corruption without modifying data.

B-Tree Index Verification

-- Lightweight check (AccessShareLock, safe for production)
SELECT bt_index_check('my_index');

-- With heap-all-indexed verification
SELECT bt_index_check('my_index', heapallindexed => true);

-- Thorough check including parent/child invariants (ShareLock, blocks writes)
SELECT bt_index_parent_check('my_index');

-- Most thorough: rootdescend re-finds each tuple from root
SELECT bt_index_parent_check('my_index',
    heapallindexed => true,
    rootdescend => true,
    checkunique => true);

Check All Catalog Indexes

SELECT bt_index_check(c.oid), c.relname, c.relpages
FROM pg_index i
JOIN pg_opclass op ON i.indclass[0] = op.oid
JOIN pg_am am ON op.opcmethod = am.oid
JOIN pg_class c ON i.indexrelid = c.oid
JOIN pg_namespace n ON c.relnamespace = n.oid
WHERE am.amname = 'btree' AND n.nspname = 'pg_catalog'
  AND c.relpersistence != 't' AND c.relkind = 'i'
  AND i.indisready AND i.indisvalid
ORDER BY c.relpages DESC LIMIT 10;

GIN Index Verification

SELECT gin_index_check('my_gin_index');

Heap (Table) Verification

-- Basic heap check
SELECT * FROM verify_heapam('my_table');

-- With TOAST verification (slower)
SELECT * FROM verify_heapam('my_table', check_toast => true);

-- Check specific block range
SELECT * FROM verify_heapam('my_table', startblock => 0, endblock => 1000);

-- Stop at first corrupted block
SELECT * FROM verify_heapam('my_table', on_error_stop => true);

Returns rows for each detected problem:

Column	Type	Description
`blkno`	bigint	Block number with corruption
`offnum`	integer	Offset of corrupt tuple
`attnum`	integer	Attribute number (if column-specific)
`msg`	text	Description of the problem

Function Summary

Function	Lock	Use Case
`bt_index_check(index, heapallindexed, checkunique)`	AccessShareLock	Routine production checks
`bt_index_parent_check(index, heapallindexed, rootdescend, checkunique)`	ShareLock	Comprehensive verification
`gin_index_check(index)`	AccessShareLock	GIN index verification
`verify_heapam(relation, on_error_stop, check_toast, skip, startblock, endblock)`	AccessShareLock	Table/heap corruption detection

All amcheck errors are true positives. Use REINDEX or point-in-time recovery for repair after diagnosing with pageinspect.

pg_surgery

Mon, 01 Jan 0001 00:00:00 +0000

pg_surgery : extension to perform surgery on a damaged relation

Overview

ID	Extension	Package	Version	Category	License	Language
5990	pg_surgery	pg_surgery	`1.0`	ADMIN	PostgreSQL	C

Attribute	Has Binary	Has Library	Need Load	Has DDL	Relocatable	Trusted
--s-d--	No	Yes	No	Yes	no	no

Relationships
See Also	pg_catcheck pg_checksums amcheck pg_orphaned pg_pathcheck pageinspect pg_repack lo lolor pg_isok

Packages

PG18	PG17	PG16	PG15	PG14
1.0	1.0	1.0	1.0	1.0

This is a built-in contrib extension ship with the PostgreSQL kernel

Install

Create this extension with:

CREATE EXTENSION pg_surgery;

Usage

pg_surgery: extension to perform surgery on a damaged relation

The pg_surgery extension provides functions to perform low-level surgery on damaged relations. These are last-resort tools that can corrupt data if misused.

Functions

heap_force_kill

Marks line pointers as “dead” without examining the tuples, forcibly removing inaccessible tuples.

heap_force_kill(regclass, tid[]) RETURNS void

-- Tuple causes error when accessed
SELECT * FROM t1 WHERE ctid = '(0, 1)';
-- ERROR: could not access status of transaction 4007513275

-- Force kill the problematic tuple
SELECT heap_force_kill('t1'::regclass, ARRAY['(0, 1)']::tid[]);

-- Tuple is now removed
SELECT * FROM t1 WHERE ctid = '(0, 1)';
-- (0 rows)

heap_force_freeze

Marks tuples as frozen without examining tuple data, making tuples accessible when visibility information is corrupted.

heap_force_freeze(regclass, tid[]) RETURNS void

-- VACUUM fails on corrupted visibility info
VACUUM t1;
-- ERROR: found xmin 507 from before relfrozenxid 515

-- Find the problematic tuple
SELECT ctid FROM t1 WHERE xmin = 507;
--  ctid
-- -------
--  (0,3)

-- Force freeze the tuple
SELECT heap_force_freeze('t1'::regclass, ARRAY['(0, 3)']::tid[]);

-- Tuple is now frozen (xmin becomes 2 = FrozenTransactionId)
SELECT ctid FROM t1 WHERE xmin = 2;
--  ctid
-- -------
--  (0,3)

When to Use

heap_force_kill: When tuples cause errors on access due to corrupted transaction status, and the data can be discarded
heap_force_freeze: When VACUUM fails due to tuples with xmin before relfrozenxid, or when tuples are invisible due to corrupted visibility information

These functions are unsafe by design and should only be used as a last resort when normal recovery methods have failed.