<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>9.7. Pattern Matching</title><link rel="stylesheet" type="text/css" href="stylesheet.css" /><link rev="made" href="pgsql-docs@lists.postgresql.org" /><meta name="generator" content="DocBook XSL Stylesheets Vsnapshot" /><link rel="prev" href="functions-bitstring.html" title="9.6. Bit String Functions and Operators" /><link rel="next" href="functions-formatting.html" title="9.8. Data Type Formatting Functions" /></head><body><div xmlns="http://www.w3.org/TR/xhtml1/transitional" class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="5" align="center">9.7. Pattern Matching</th></tr><tr><td width="10%" align="left"><a accesskey="p" href="functions-bitstring.html" title="9.6. Bit String Functions and Operators">Prev</a> </td><td width="10%" align="left"><a accesskey="u" href="functions.html" title="Chapter 9. Functions and Operators">Up</a></td><th width="60%" align="center">Chapter 9. Functions and Operators</th><td width="10%" align="right"><a accesskey="h" href="index.html" title="PostgreSQL 11.5 Documentation">Home</a></td><td width="10%" align="right"> <a accesskey="n" href="functions-formatting.html" title="9.8. Data Type Formatting Functions">Next</a></td></tr></table><hr></hr></div><div class="sect1" id="FUNCTIONS-MATCHING"><div class="titlepage"><div><div><h2 class="title" style="clear: both">9.7. Pattern Matching</h2></div></div></div><div class="toc"><dl class="toc"><dt><span class="sect2"><a href="functions-matching.html#FUNCTIONS-LIKE">9.7.1. <code class="function">LIKE</code></a></span></dt><dt><span class="sect2"><a href="functions-matching.html#FUNCTIONS-SIMILARTO-REGEXP">9.7.2. <code class="function">SIMILAR TO</code> Regular Expressions</a></span></dt><dt><span class="sect2"><a href="functions-matching.html#FUNCTIONS-POSIX-REGEXP">9.7.3. <acronym class="acronym">POSIX</acronym> Regular Expressions</a></span></dt></dl></div><a id="id-1.5.8.12.2" class="indexterm"></a><p>
There are three separate approaches to pattern matching provided
by <span class="productname">PostgreSQL</span>: the traditional
<acronym class="acronym">SQL</acronym> <code class="function">LIKE</code> operator, the
more recent <code class="function">SIMILAR TO</code> operator (added in
SQL:1999), and <acronym class="acronym">POSIX</acronym>-style regular
expressions. Aside from the basic <span class="quote">“<span class="quote">does this string match
this pattern?</span>”</span> operators, functions are available to extract
or replace matching substrings and to split a string at matching
locations.
</p><div class="tip"><h3 class="title">Tip</h3><p>
If you have pattern matching needs that go beyond this,
consider writing a user-defined function in Perl or Tcl.
</p></div><div class="caution"><h3 class="title">Caution</h3><p>
While most regular-expression searches can be executed very quickly,
regular expressions can be contrived that take arbitrary amounts of
time and memory to process. Be wary of accepting regular-expression
search patterns from hostile sources. If you must do so, it is
advisable to impose a statement timeout.
</p><p>
Searches using <code class="function">SIMILAR TO</code> patterns have the same
security hazards, since <code class="function">SIMILAR TO</code> provides many
of the same capabilities as <acronym class="acronym">POSIX</acronym>-style regular
expressions.
</p><p>
<code class="function">LIKE</code> searches, being much simpler than the other
two options, are safer to use with possibly-hostile pattern sources.
</p></div><div class="sect2" id="FUNCTIONS-LIKE"><div class="titlepage"><div><div><h3 class="title">9.7.1. <code class="function">LIKE</code></h3></div></div></div><a id="id-1.5.8.12.6.2" class="indexterm"></a><pre class="synopsis">
<em class="replaceable"><code>string</code></em> LIKE <em class="replaceable"><code>pattern</code></em> [<span class="optional">ESCAPE <em class="replaceable"><code>escape-character</code></em></span>]
<em class="replaceable"><code>string</code></em> NOT LIKE <em class="replaceable"><code>pattern</code></em> [<span class="optional">ESCAPE <em class="replaceable"><code>escape-character</code></em></span>]
</pre><p>
The <code class="function">LIKE</code> expression returns true if the
<em class="replaceable"><code>string</code></em> matches the supplied
<em class="replaceable"><code>pattern</code></em>. (As
expected, the <code class="function">NOT LIKE</code> expression returns
false if <code class="function">LIKE</code> returns true, and vice versa.
An equivalent expression is
<code class="literal">NOT (<em class="replaceable"><code>string</code></em> LIKE
<em class="replaceable"><code>pattern</code></em>)</code>.)
</p><p>
If <em class="replaceable"><code>pattern</code></em> does not contain percent
signs or underscores, then the pattern only represents the string
itself; in that case <code class="function">LIKE</code> acts like the
equals operator. An underscore (<code class="literal">_</code>) in
<em class="replaceable"><code>pattern</code></em> stands for (matches) any single
character; a percent sign (<code class="literal">%</code>) matches any sequence
of zero or more characters.
</p><p>
Some examples:
</p><pre class="programlisting">
'abc' LIKE 'abc' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' LIKE 'a%' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' LIKE '_b_' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' LIKE 'c' <em class="lineannotation"><span class="lineannotation">false</span></em>
</pre><p>
</p><p>
<code class="function">LIKE</code> pattern matching always covers the entire
string. Therefore, if it's desired to match a sequence anywhere within
a string, the pattern must start and end with a percent sign.
</p><p>
To match a literal underscore or percent sign without matching
other characters, the respective character in
<em class="replaceable"><code>pattern</code></em> must be
preceded by the escape character. The default escape
character is the backslash but a different one can be selected by
using the <code class="literal">ESCAPE</code> clause. To match the escape
character itself, write two escape characters.
</p><div class="note"><h3 class="title">Note</h3><p>
If you have <a class="xref" href="runtime-config-compatible.html#GUC-STANDARD-CONFORMING-STRINGS">standard_conforming_strings</a> turned off,
any backslashes you write in literal string constants will need to be
doubled. See <a class="xref" href="sql-syntax-lexical.html#SQL-SYNTAX-STRINGS" title="4.1.2.1. String Constants">Section 4.1.2.1</a> for more information.
</p></div><p>
It's also possible to select no escape character by writing
<code class="literal">ESCAPE ''</code>. This effectively disables the
escape mechanism, which makes it impossible to turn off the
special meaning of underscore and percent signs in the pattern.
</p><p>
The key word <code class="token">ILIKE</code> can be used instead of
<code class="token">LIKE</code> to make the match case-insensitive according
to the active locale. This is not in the <acronym class="acronym">SQL</acronym> standard but is a
<span class="productname">PostgreSQL</span> extension.
</p><p>
The operator <code class="literal">~~</code> is equivalent to
<code class="function">LIKE</code>, and <code class="literal">~~*</code> corresponds to
<code class="function">ILIKE</code>. There are also
<code class="literal">!~~</code> and <code class="literal">!~~*</code> operators that
represent <code class="function">NOT LIKE</code> and <code class="function">NOT
ILIKE</code>, respectively. All of these operators are
<span class="productname">PostgreSQL</span>-specific.
</p><p>
There is also the prefix operator <code class="literal">^@</code> and corresponding
<code class="function">starts_with</code> function which covers cases when only
searching by beginning of the string is needed.
</p></div><div class="sect2" id="FUNCTIONS-SIMILARTO-REGEXP"><div class="titlepage"><div><div><h3 class="title">9.7.2. <code class="function">SIMILAR TO</code> Regular Expressions</h3></div></div></div><a id="id-1.5.8.12.7.2" class="indexterm"></a><a id="id-1.5.8.12.7.3" class="indexterm"></a><a id="id-1.5.8.12.7.4" class="indexterm"></a><pre class="synopsis">
<em class="replaceable"><code>string</code></em> SIMILAR TO <em class="replaceable"><code>pattern</code></em> [<span class="optional">ESCAPE <em class="replaceable"><code>escape-character</code></em></span>]
<em class="replaceable"><code>string</code></em> NOT SIMILAR TO <em class="replaceable"><code>pattern</code></em> [<span class="optional">ESCAPE <em class="replaceable"><code>escape-character</code></em></span>]
</pre><p>
The <code class="function">SIMILAR TO</code> operator returns true or
false depending on whether its pattern matches the given string.
It is similar to <code class="function">LIKE</code>, except that it
interprets the pattern using the SQL standard's definition of a
regular expression. SQL regular expressions are a curious cross
between <code class="function">LIKE</code> notation and common regular
expression notation.
</p><p>
Like <code class="function">LIKE</code>, the <code class="function">SIMILAR TO</code>
operator succeeds only if its pattern matches the entire string;
this is unlike common regular expression behavior where the pattern
can match any part of the string.
Also like
<code class="function">LIKE</code>, <code class="function">SIMILAR TO</code> uses
<code class="literal">_</code> and <code class="literal">%</code> as wildcard characters denoting
any single character and any string, respectively (these are
comparable to <code class="literal">.</code> and <code class="literal">.*</code> in POSIX regular
expressions).
</p><p>
In addition to these facilities borrowed from <code class="function">LIKE</code>,
<code class="function">SIMILAR TO</code> supports these pattern-matching
metacharacters borrowed from POSIX regular expressions:
</p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
<code class="literal">|</code> denotes alternation (either of two alternatives).
</p></li><li class="listitem"><p>
<code class="literal">*</code> denotes repetition of the previous item zero
or more times.
</p></li><li class="listitem"><p>
<code class="literal">+</code> denotes repetition of the previous item one
or more times.
</p></li><li class="listitem"><p>
<code class="literal">?</code> denotes repetition of the previous item zero
or one time.
</p></li><li class="listitem"><p>
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">}</code> denotes repetition
of the previous item exactly <em class="replaceable"><code>m</code></em> times.
</p></li><li class="listitem"><p>
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,}</code> denotes repetition
of the previous item <em class="replaceable"><code>m</code></em> or more times.
</p></li><li class="listitem"><p>
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,</code><em class="replaceable"><code>n</code></em><code class="literal">}</code>
denotes repetition of the previous item at least <em class="replaceable"><code>m</code></em> and
not more than <em class="replaceable"><code>n</code></em> times.
</p></li><li class="listitem"><p>
Parentheses <code class="literal">()</code> can be used to group items into
a single logical item.
</p></li><li class="listitem"><p>
A bracket expression <code class="literal">[...]</code> specifies a character
class, just as in POSIX regular expressions.
</p></li></ul></div><p>
Notice that the period (<code class="literal">.</code>) is not a metacharacter
for <code class="function">SIMILAR TO</code>.
</p><p>
As with <code class="function">LIKE</code>, a backslash disables the special meaning
of any of these metacharacters; or a different escape character can
be specified with <code class="literal">ESCAPE</code>.
</p><p>
Some examples:
</p><pre class="programlisting">
'abc' SIMILAR TO 'abc' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' SIMILAR TO 'a' <em class="lineannotation"><span class="lineannotation">false</span></em>
'abc' SIMILAR TO '%(b|d)%' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' SIMILAR TO '(b|c)%' <em class="lineannotation"><span class="lineannotation">false</span></em>
</pre><p>
</p><p>
The <code class="function">substring</code> function with three parameters,
<code class="function">substring(<em class="replaceable"><code>string</code></em> from
<em class="replaceable"><code>pattern</code></em> for
<em class="replaceable"><code>escape-character</code></em>)</code>, provides
extraction of a substring that matches an SQL
regular expression pattern. As with <code class="literal">SIMILAR TO</code>, the
specified pattern must match the entire data string, or else the
function fails and returns null. To indicate the part of the
pattern that should be returned on success, the pattern must contain
two occurrences of the escape character followed by a double quote
(<code class="literal">"</code>).
The text matching the portion of the pattern
between these markers is returned.
</p><p>
Some examples, with <code class="literal">#"</code> delimiting the return string:
</p><pre class="programlisting">
substring('foobar' from '%#"o_b#"%' for '#') <em class="lineannotation"><span class="lineannotation">oob</span></em>
substring('foobar' from '#"o_b#"%' for '#') <em class="lineannotation"><span class="lineannotation">NULL</span></em>
</pre><p>
</p></div><div class="sect2" id="FUNCTIONS-POSIX-REGEXP"><div class="titlepage"><div><div><h3 class="title">9.7.3. <acronym class="acronym">POSIX</acronym> Regular Expressions</h3></div></div></div><a id="id-1.5.8.12.8.2" class="indexterm"></a><a id="id-1.5.8.12.8.3" class="indexterm"></a><a id="id-1.5.8.12.8.4" class="indexterm"></a><a id="id-1.5.8.12.8.5" class="indexterm"></a><a id="id-1.5.8.12.8.6" class="indexterm"></a><a id="id-1.5.8.12.8.7" class="indexterm"></a><a id="id-1.5.8.12.8.8" class="indexterm"></a><p>
<a class="xref" href="functions-matching.html#FUNCTIONS-POSIX-TABLE" title="Table 9.14. Regular Expression Match Operators">Table 9.14</a> lists the available
operators for pattern matching using POSIX regular expressions.
</p><div class="table" id="FUNCTIONS-POSIX-TABLE"><p class="title"><strong>Table 9.14. Regular Expression Match Operators</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Match Operators" border="1"><colgroup><col /><col /><col /></colgroup><thead><tr><th>Operator</th><th>Description</th><th>Example</th></tr></thead><tbody><tr><td> <code class="literal">~</code> </td><td>Matches regular expression, case sensitive</td><td><code class="literal">'thomas' ~ '.*thomas.*'</code></td></tr><tr><td> <code class="literal">~*</code> </td><td>Matches regular expression, case insensitive</td><td><code class="literal">'thomas' ~* '.*Thomas.*'</code></td></tr><tr><td> <code class="literal">!~</code> </td><td>Does not match regular expression, case sensitive</td><td><code class="literal">'thomas' !~ '.*Thomas.*'</code></td></tr><tr><td> <code class="literal">!~*</code> </td><td>Does not match regular expression, case insensitive</td><td><code class="literal">'thomas' !~* '.*vadim.*'</code></td></tr></tbody></table></div></div><br class="table-break" /><p>
<acronym class="acronym">POSIX</acronym> regular expressions provide a more
powerful means for pattern matching than the <code class="function">LIKE</code> and
<code class="function">SIMILAR TO</code> operators.
Many Unix tools such as <code class="command">egrep</code>,
<code class="command">sed</code>, or <code class="command">awk</code> use a pattern
matching language that is similar to the one described here.
</p><p>
A regular expression is a character sequence that is an
abbreviated definition of a set of strings (a <em class="firstterm">regular
set</em>). A string is said to match a regular expression
if it is a member of the regular set described by the regular
expression. As with <code class="function">LIKE</code>, pattern characters
match string characters exactly unless they are special characters
in the regular expression language — but regular expressions use
different special characters than <code class="function">LIKE</code> does.
Unlike <code class="function">LIKE</code> patterns, a
regular expression is allowed to match anywhere within a string, unless
the regular expression is explicitly anchored to the beginning or
end of the string.
</p><p>
Some examples:
</p><pre class="programlisting">
'abc' ~ 'abc' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' ~ '^a' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' ~ '(b|d)' <em class="lineannotation"><span class="lineannotation">true</span></em>
'abc' ~ '^(b|c)' <em class="lineannotation"><span class="lineannotation">false</span></em>
</pre><p>
</p><p>
The <acronym class="acronym">POSIX</acronym> pattern language is described in much
greater detail below.
</p><p>
The <code class="function">substring</code> function with two parameters,
<code class="function">substring(<em class="replaceable"><code>string</code></em> from
<em class="replaceable"><code>pattern</code></em>)</code>, provides extraction of a
substring
that matches a POSIX regular expression pattern. It returns null if
there is no match, otherwise the portion of the text that matched the
pattern. But if the pattern contains any parentheses, the portion
of the text that matched the first parenthesized subexpression (the
one whose left parenthesis comes first) is
returned. You can put parentheses around the whole expression
if you want to use parentheses within it without triggering this
exception. If you need parentheses in the pattern before the
subexpression you want to extract, see the non-capturing parentheses
described below.
</p><p>
Some examples:
</p><pre class="programlisting">
substring('foobar' from 'o.b') <em class="lineannotation"><span class="lineannotation">oob</span></em>
substring('foobar' from 'o(.)b') <em class="lineannotation"><span class="lineannotation">o</span></em>
</pre><p>
</p><p>
The <code class="function">regexp_replace</code> function provides substitution of
new text for substrings that match POSIX regular expression patterns.
It has the syntax
<code class="function">regexp_replace</code>(<em class="replaceable"><code>source</code></em>,
<em class="replaceable"><code>pattern</code></em>, <em class="replaceable"><code>replacement</code></em>
[<span class="optional">, <em class="replaceable"><code>flags</code></em> </span>]).
The <em class="replaceable"><code>source</code></em> string is returned unchanged if
there is no match to the <em class="replaceable"><code>pattern</code></em>. If there is a
match, the <em class="replaceable"><code>source</code></em> string is returned with the
<em class="replaceable"><code>replacement</code></em> string substituted for the matching
substring. The <em class="replaceable"><code>replacement</code></em> string can contain
<code class="literal">\</code><em class="replaceable"><code>n</code></em>, where <em class="replaceable"><code>n</code></em> is 1
through 9, to indicate that the source substring matching the
<em class="replaceable"><code>n</code></em>'th parenthesized subexpression of the pattern should be
inserted, and it can contain <code class="literal">\&</code> to indicate that the
substring matching the entire pattern should be inserted. Write
<code class="literal">\\</code> if you need to put a literal backslash in the replacement
text.
The <em class="replaceable"><code>flags</code></em> parameter is an optional text
string containing zero or more single-letter flags that change the
function's behavior. Flag <code class="literal">i</code> specifies case-insensitive
matching, while flag <code class="literal">g</code> specifies replacement of each matching
substring rather than only the first one. Supported flags (though
not <code class="literal">g</code>) are
described in <a class="xref" href="functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE" title="Table 9.22. ARE Embedded-option Letters">Table 9.22</a>.
</p><p>
Some examples:
</p><pre class="programlisting">
regexp_replace('foobarbaz', 'b..', 'X')
<em class="lineannotation"><span class="lineannotation">fooXbaz</span></em>
regexp_replace('foobarbaz', 'b..', 'X', 'g')
<em class="lineannotation"><span class="lineannotation">fooXX</span></em>
regexp_replace('foobarbaz', 'b(..)', 'X\1Y', 'g')
<em class="lineannotation"><span class="lineannotation">fooXarYXazY</span></em>
</pre><p>
</p><p>
The <code class="function">regexp_match</code> function returns a text array of
captured substring(s) resulting from the first match of a POSIX
regular expression pattern to a string. It has the syntax
<code class="function">regexp_match</code>(<em class="replaceable"><code>string</code></em>,
<em class="replaceable"><code>pattern</code></em> [<span class="optional">, <em class="replaceable"><code>flags</code></em> </span>]).
If there is no match, the result is <code class="literal">NULL</code>.
If a match is found, and the <em class="replaceable"><code>pattern</code></em> contains no
parenthesized subexpressions, then the result is a single-element text
array containing the substring matching the whole pattern.
If a match is found, and the <em class="replaceable"><code>pattern</code></em> contains
parenthesized subexpressions, then the result is a text array
whose <em class="replaceable"><code>n</code></em>'th element is the substring matching
the <em class="replaceable"><code>n</code></em>'th parenthesized subexpression of
the <em class="replaceable"><code>pattern</code></em> (not counting <span class="quote">“<span class="quote">non-capturing</span>”</span>
parentheses; see below for details).
The <em class="replaceable"><code>flags</code></em> parameter is an optional text string
containing zero or more single-letter flags that change the function's
behavior. Supported flags are described
in <a class="xref" href="functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE" title="Table 9.22. ARE Embedded-option Letters">Table 9.22</a>.
</p><p>
Some examples:
</p><pre class="programlisting">
SELECT regexp_match('foobarbequebaz', 'bar.*que');
regexp_match
--------------
{barbeque}
(1 row)
SELECT regexp_match('foobarbequebaz', '(bar)(beque)');
regexp_match
--------------
{bar,beque}
(1 row)
</pre><p>
In the common case where you just want the whole matching substring
or <code class="literal">NULL</code> for no match, write something like
</p><pre class="programlisting">
SELECT (regexp_match('foobarbequebaz', 'bar.*que'))[1];
regexp_match
--------------
barbeque
(1 row)
</pre><p>
</p><p>
The <code class="function">regexp_matches</code> function returns a set of text arrays
of captured substring(s) resulting from matching a POSIX regular
expression pattern to a string. It has the same syntax as
<code class="function">regexp_match</code>.
This function returns no rows if there is no match, one row if there is
a match and the <code class="literal">g</code> flag is not given, or <em class="replaceable"><code>N</code></em>
rows if there are <em class="replaceable"><code>N</code></em> matches and the <code class="literal">g</code> flag
is given. Each returned row is a text array containing the whole
matched substring or the substrings matching parenthesized
subexpressions of the <em class="replaceable"><code>pattern</code></em>, just as described above
for <code class="function">regexp_match</code>.
<code class="function">regexp_matches</code> accepts all the flags shown
in <a class="xref" href="functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE" title="Table 9.22. ARE Embedded-option Letters">Table 9.22</a>, plus
the <code class="literal">g</code> flag which commands it to return all matches, not
just the first one.
</p><p>
Some examples:
</p><pre class="programlisting">
SELECT regexp_matches('foo', 'not there');
regexp_matches
----------------
(0 rows)
SELECT regexp_matches('foobarbequebazilbarfbonk', '(b[^b]+)(b[^b]+)', 'g');
regexp_matches
----------------
{bar,beque}
{bazil,barf}
(2 rows)
</pre><p>
</p><div class="tip"><h3 class="title">Tip</h3><p>
In most cases <code class="function">regexp_matches()</code> should be used with
the <code class="literal">g</code> flag, since if you only want the first match, it's
easier and more efficient to use <code class="function">regexp_match()</code>.
However, <code class="function">regexp_match()</code> only exists
in <span class="productname">PostgreSQL</span> version 10 and up. When working in older
versions, a common trick is to place a <code class="function">regexp_matches()</code>
call in a sub-select, for example:
</p><pre class="programlisting">
SELECT col1, (SELECT regexp_matches(col2, '(bar)(beque)')) FROM tab;
</pre><p>
This produces a text array if there's a match, or <code class="literal">NULL</code> if
not, the same as <code class="function">regexp_match()</code> would do. Without the
sub-select, this query would produce no output at all for table rows
without a match, which is typically not the desired behavior.
</p></div><p>
The <code class="function">regexp_split_to_table</code> function splits a string using a POSIX
regular expression pattern as a delimiter. It has the syntax
<code class="function">regexp_split_to_table</code>(<em class="replaceable"><code>string</code></em>, <em class="replaceable"><code>pattern</code></em>
[<span class="optional">, <em class="replaceable"><code>flags</code></em> </span>]).
If there is no match to the <em class="replaceable"><code>pattern</code></em>, the function returns the
<em class="replaceable"><code>string</code></em>. If there is at least one match, for each match it returns
the text from the end of the last match (or the beginning of the string)
to the beginning of the match. When there are no more matches, it
returns the text from the end of the last match to the end of the string.
The <em class="replaceable"><code>flags</code></em> parameter is an optional text string containing
zero or more single-letter flags that change the function's behavior.
<code class="function">regexp_split_to_table</code> supports the flags described in
<a class="xref" href="functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE" title="Table 9.22. ARE Embedded-option Letters">Table 9.22</a>.
</p><p>
The <code class="function">regexp_split_to_array</code> function behaves the same as
<code class="function">regexp_split_to_table</code>, except that <code class="function">regexp_split_to_array</code>
returns its result as an array of <code class="type">text</code>. It has the syntax
<code class="function">regexp_split_to_array</code>(<em class="replaceable"><code>string</code></em>, <em class="replaceable"><code>pattern</code></em>
[<span class="optional">, <em class="replaceable"><code>flags</code></em> </span>]).
The parameters are the same as for <code class="function">regexp_split_to_table</code>.
</p><p>
Some examples:
</p><pre class="programlisting">
SELECT foo FROM regexp_split_to_table('the quick brown fox jumps over the lazy dog', '\s+') AS foo;
foo
-------
the
quick
brown
fox
jumps
over
the
lazy
dog
(9 rows)
SELECT regexp_split_to_array('the quick brown fox jumps over the lazy dog', '\s+');
regexp_split_to_array
-----------------------------------------------
{the,quick,brown,fox,jumps,over,the,lazy,dog}
(1 row)
SELECT foo FROM regexp_split_to_table('the quick brown fox', '\s*') AS foo;
foo
-----
t
h
e
q
u
i
c
k
b
r
o
w
n
f
o
x
(16 rows)
</pre><p>
</p><p>
As the last example demonstrates, the regexp split functions ignore
zero-length matches that occur at the start or end of the string
or immediately after a previous match. This is contrary to the strict
definition of regexp matching that is implemented by
<code class="function">regexp_match</code> and
<code class="function">regexp_matches</code>, but is usually the most convenient behavior
in practice. Other software systems such as Perl use similar definitions.
</p><div class="sect3" id="POSIX-SYNTAX-DETAILS"><div class="titlepage"><div><div><h4 class="title">9.7.3.1. Regular Expression Details</h4></div></div></div><p>
<span class="productname">PostgreSQL</span>'s regular expressions are implemented
using a software package written by Henry Spencer. Much of
the description of regular expressions below is copied verbatim from his
manual.
</p><p>
Regular expressions (<acronym class="acronym">RE</acronym>s), as defined in
<acronym class="acronym">POSIX</acronym> 1003.2, come in two forms:
<em class="firstterm">extended</em> <acronym class="acronym">RE</acronym>s or <acronym class="acronym">ERE</acronym>s
(roughly those of <code class="command">egrep</code>), and
<em class="firstterm">basic</em> <acronym class="acronym">RE</acronym>s or <acronym class="acronym">BRE</acronym>s
(roughly those of <code class="command">ed</code>).
<span class="productname">PostgreSQL</span> supports both forms, and
also implements some extensions
that are not in the POSIX standard, but have become widely used
due to their availability in programming languages such as Perl and Tcl.
<acronym class="acronym">RE</acronym>s using these non-POSIX extensions are called
<em class="firstterm">advanced</em> <acronym class="acronym">RE</acronym>s or <acronym class="acronym">ARE</acronym>s
in this documentation. AREs are almost an exact superset of EREs,
but BREs have several notational incompatibilities (as well as being
much more limited).
We first describe the ARE and ERE forms, noting features that apply
only to AREs, and then describe how BREs differ.
</p><div class="note"><h3 class="title">Note</h3><p>
<span class="productname">PostgreSQL</span> always initially presumes that a regular
expression follows the ARE rules. However, the more limited ERE or
BRE rules can be chosen by prepending an <em class="firstterm">embedded option</em>
to the RE pattern, as described in <a class="xref" href="functions-matching.html#POSIX-METASYNTAX" title="9.7.3.4. Regular Expression Metasyntax">Section 9.7.3.4</a>.
This can be useful for compatibility with applications that expect
exactly the <acronym class="acronym">POSIX</acronym> 1003.2 rules.
</p></div><p>
A regular expression is defined as one or more
<em class="firstterm">branches</em>, separated by
<code class="literal">|</code>. It matches anything that matches one of the
branches.
</p><p>
A branch is zero or more <em class="firstterm">quantified atoms</em> or
<em class="firstterm">constraints</em>, concatenated.
It matches a match for the first, followed by a match for the second, etc;
an empty branch matches the empty string.
</p><p>
A quantified atom is an <em class="firstterm">atom</em> possibly followed
by a single <em class="firstterm">quantifier</em>.
Without a quantifier, it matches a match for the atom.
With a quantifier, it can match some number of matches of the atom.
An <em class="firstterm">atom</em> can be any of the possibilities
shown in <a class="xref" href="functions-matching.html#POSIX-ATOMS-TABLE" title="Table 9.15. Regular Expression Atoms">Table 9.15</a>.
The possible quantifiers and their meanings are shown in
<a class="xref" href="functions-matching.html#POSIX-QUANTIFIERS-TABLE" title="Table 9.16. Regular Expression Quantifiers">Table 9.16</a>.
</p><p>
A <em class="firstterm">constraint</em> matches an empty string, but matches only when
specific conditions are met. A constraint can be used where an atom
could be used, except it cannot be followed by a quantifier.
The simple constraints are shown in
<a class="xref" href="functions-matching.html#POSIX-CONSTRAINTS-TABLE" title="Table 9.17. Regular Expression Constraints">Table 9.17</a>;
some more constraints are described later.
</p><div class="table" id="POSIX-ATOMS-TABLE"><p class="title"><strong>Table 9.15. Regular Expression Atoms</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Atoms" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Atom</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">(</code><em class="replaceable"><code>re</code></em><code class="literal">)</code> </td><td> (where <em class="replaceable"><code>re</code></em> is any regular expression)
matches a match for
<em class="replaceable"><code>re</code></em>, with the match noted for possible reporting </td></tr><tr><td> <code class="literal">(?:</code><em class="replaceable"><code>re</code></em><code class="literal">)</code> </td><td> as above, but the match is not noted for reporting
(a <span class="quote">“<span class="quote">non-capturing</span>”</span> set of parentheses)
(AREs only) </td></tr><tr><td> <code class="literal">.</code> </td><td> matches any single character </td></tr><tr><td> <code class="literal">[</code><em class="replaceable"><code>chars</code></em><code class="literal">]</code> </td><td> a <em class="firstterm">bracket expression</em>,
matching any one of the <em class="replaceable"><code>chars</code></em> (see
<a class="xref" href="functions-matching.html#POSIX-BRACKET-EXPRESSIONS" title="9.7.3.2. Bracket Expressions">Section 9.7.3.2</a> for more detail) </td></tr><tr><td> <code class="literal">\</code><em class="replaceable"><code>k</code></em> </td><td> (where <em class="replaceable"><code>k</code></em> is a non-alphanumeric character)
matches that character taken as an ordinary character,
e.g., <code class="literal">\\</code> matches a backslash character </td></tr><tr><td> <code class="literal">\</code><em class="replaceable"><code>c</code></em> </td><td> where <em class="replaceable"><code>c</code></em> is alphanumeric
(possibly followed by other characters)
is an <em class="firstterm">escape</em>, see <a class="xref" href="functions-matching.html#POSIX-ESCAPE-SEQUENCES" title="9.7.3.3. Regular Expression Escapes">Section 9.7.3.3</a>
(AREs only; in EREs and BREs, this matches <em class="replaceable"><code>c</code></em>) </td></tr><tr><td> <code class="literal">{</code> </td><td> when followed by a character other than a digit,
matches the left-brace character <code class="literal">{</code>;
when followed by a digit, it is the beginning of a
<em class="replaceable"><code>bound</code></em> (see below) </td></tr><tr><td> <em class="replaceable"><code>x</code></em> </td><td> where <em class="replaceable"><code>x</code></em> is a single character with no other
significance, matches that character </td></tr></tbody></table></div></div><br class="table-break" /><p>
An RE cannot end with a backslash (<code class="literal">\</code>).
</p><div class="note"><h3 class="title">Note</h3><p>
If you have <a class="xref" href="runtime-config-compatible.html#GUC-STANDARD-CONFORMING-STRINGS">standard_conforming_strings</a> turned off,
any backslashes you write in literal string constants will need to be
doubled. See <a class="xref" href="sql-syntax-lexical.html#SQL-SYNTAX-STRINGS" title="4.1.2.1. String Constants">Section 4.1.2.1</a> for more information.
</p></div><div class="table" id="POSIX-QUANTIFIERS-TABLE"><p class="title"><strong>Table 9.16. Regular Expression Quantifiers</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Quantifiers" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Quantifier</th><th>Matches</th></tr></thead><tbody><tr><td> <code class="literal">*</code> </td><td> a sequence of 0 or more matches of the atom </td></tr><tr><td> <code class="literal">+</code> </td><td> a sequence of 1 or more matches of the atom </td></tr><tr><td> <code class="literal">?</code> </td><td> a sequence of 0 or 1 matches of the atom </td></tr><tr><td> <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">}</code> </td><td> a sequence of exactly <em class="replaceable"><code>m</code></em> matches of the atom </td></tr><tr><td> <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,}</code> </td><td> a sequence of <em class="replaceable"><code>m</code></em> or more matches of the atom </td></tr><tr><td>
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,</code><em class="replaceable"><code>n</code></em><code class="literal">}</code> </td><td> a sequence of <em class="replaceable"><code>m</code></em> through <em class="replaceable"><code>n</code></em>
(inclusive) matches of the atom; <em class="replaceable"><code>m</code></em> cannot exceed
<em class="replaceable"><code>n</code></em> </td></tr><tr><td> <code class="literal">*?</code> </td><td> non-greedy version of <code class="literal">*</code> </td></tr><tr><td> <code class="literal">+?</code> </td><td> non-greedy version of <code class="literal">+</code> </td></tr><tr><td> <code class="literal">??</code> </td><td> non-greedy version of <code class="literal">?</code> </td></tr><tr><td> <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">}?</code> </td><td> non-greedy version of <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">}</code> </td></tr><tr><td> <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,}?</code> </td><td> non-greedy version of <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,}</code> </td></tr><tr><td>
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,</code><em class="replaceable"><code>n</code></em><code class="literal">}?</code> </td><td> non-greedy version of <code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,</code><em class="replaceable"><code>n</code></em><code class="literal">}</code> </td></tr></tbody></table></div></div><br class="table-break" /><p>
The forms using <code class="literal">{</code><em class="replaceable"><code>...</code></em><code class="literal">}</code>
are known as <em class="firstterm">bounds</em>.
The numbers <em class="replaceable"><code>m</code></em> and <em class="replaceable"><code>n</code></em> within a bound are
unsigned decimal integers with permissible values from 0 to 255 inclusive.
</p><p>
<em class="firstterm">Non-greedy</em> quantifiers (available in AREs only) match the
same possibilities as their corresponding normal (<em class="firstterm">greedy</em>)
counterparts, but prefer the smallest number rather than the largest
number of matches.
See <a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a> for more detail.
</p><div class="note"><h3 class="title">Note</h3><p>
A quantifier cannot immediately follow another quantifier, e.g.,
<code class="literal">**</code> is invalid.
A quantifier cannot
begin an expression or subexpression or follow
<code class="literal">^</code> or <code class="literal">|</code>.
</p></div><div class="table" id="POSIX-CONSTRAINTS-TABLE"><p class="title"><strong>Table 9.17. Regular Expression Constraints</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Constraints" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Constraint</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">^</code> </td><td> matches at the beginning of the string </td></tr><tr><td> <code class="literal">$</code> </td><td> matches at the end of the string </td></tr><tr><td> <code class="literal">(?=</code><em class="replaceable"><code>re</code></em><code class="literal">)</code> </td><td> <em class="firstterm">positive lookahead</em> matches at any point
where a substring matching <em class="replaceable"><code>re</code></em> begins
(AREs only) </td></tr><tr><td> <code class="literal">(?!</code><em class="replaceable"><code>re</code></em><code class="literal">)</code> </td><td> <em class="firstterm">negative lookahead</em> matches at any point
where no substring matching <em class="replaceable"><code>re</code></em> begins
(AREs only) </td></tr><tr><td> <code class="literal">(?<=</code><em class="replaceable"><code>re</code></em><code class="literal">)</code> </td><td> <em class="firstterm">positive lookbehind</em> matches at any point
where a substring matching <em class="replaceable"><code>re</code></em> ends
(AREs only) </td></tr><tr><td> <code class="literal">(?<!</code><em class="replaceable"><code>re</code></em><code class="literal">)</code> </td><td> <em class="firstterm">negative lookbehind</em> matches at any point
where no substring matching <em class="replaceable"><code>re</code></em> ends
(AREs only) </td></tr></tbody></table></div></div><br class="table-break" /><p>
Lookahead and lookbehind constraints cannot contain <em class="firstterm">back
references</em> (see <a class="xref" href="functions-matching.html#POSIX-ESCAPE-SEQUENCES" title="9.7.3.3. Regular Expression Escapes">Section 9.7.3.3</a>),
and all parentheses within them are considered non-capturing.
</p></div><div class="sect3" id="POSIX-BRACKET-EXPRESSIONS"><div class="titlepage"><div><div><h4 class="title">9.7.3.2. Bracket Expressions</h4></div></div></div><p>
A <em class="firstterm">bracket expression</em> is a list of
characters enclosed in <code class="literal">[]</code>. It normally matches
any single character from the list (but see below). If the list
begins with <code class="literal">^</code>, it matches any single character
<span class="emphasis"><em>not</em></span> from the rest of the list.
If two characters
in the list are separated by <code class="literal">-</code>, this is
shorthand for the full range of characters between those two
(inclusive) in the collating sequence,
e.g., <code class="literal">[0-9]</code> in <acronym class="acronym">ASCII</acronym> matches
any decimal digit. It is illegal for two ranges to share an
endpoint, e.g., <code class="literal">a-c-e</code>. Ranges are very
collating-sequence-dependent, so portable programs should avoid
relying on them.
</p><p>
To include a literal <code class="literal">]</code> in the list, make it the
first character (after <code class="literal">^</code>, if that is used). To
include a literal <code class="literal">-</code>, make it the first or last
character, or the second endpoint of a range. To use a literal
<code class="literal">-</code> as the first endpoint of a range, enclose it
in <code class="literal">[.</code> and <code class="literal">.]</code> to make it a
collating element (see below). With the exception of these characters,
some combinations using <code class="literal">[</code>
(see next paragraphs), and escapes (AREs only), all other special
characters lose their special significance within a bracket expression.
In particular, <code class="literal">\</code> is not special when following
ERE or BRE rules, though it is special (as introducing an escape)
in AREs.
</p><p>
Within a bracket expression, a collating element (a character, a
multiple-character sequence that collates as if it were a single
character, or a collating-sequence name for either) enclosed in
<code class="literal">[.</code> and <code class="literal">.]</code> stands for the
sequence of characters of that collating element. The sequence is
treated as a single element of the bracket expression's list. This
allows a bracket
expression containing a multiple-character collating element to
match more than one character, e.g., if the collating sequence
includes a <code class="literal">ch</code> collating element, then the RE
<code class="literal">[[.ch.]]*c</code> matches the first five characters of
<code class="literal">chchcc</code>.
</p><div class="note"><h3 class="title">Note</h3><p>
<span class="productname">PostgreSQL</span> currently does not support multi-character collating
elements. This information describes possible future behavior.
</p></div><p>
Within a bracket expression, a collating element enclosed in
<code class="literal">[=</code> and <code class="literal">=]</code> is an <em class="firstterm">equivalence
class</em>, standing for the sequences of characters of all collating
elements equivalent to that one, including itself. (If there are
no other equivalent collating elements, the treatment is as if the
enclosing delimiters were <code class="literal">[.</code> and
<code class="literal">.]</code>.) For example, if <code class="literal">o</code> and
<code class="literal">^</code> are the members of an equivalence class, then
<code class="literal">[[=o=]]</code>, <code class="literal">[[=^=]]</code>, and
<code class="literal">[o^]</code> are all synonymous. An equivalence class
cannot be an endpoint of a range.
</p><p>
Within a bracket expression, the name of a character class
enclosed in <code class="literal">[:</code> and <code class="literal">:]</code> stands
for the list of all characters belonging to that class. Standard
character class names are: <code class="literal">alnum</code>,
<code class="literal">alpha</code>, <code class="literal">blank</code>,
<code class="literal">cntrl</code>, <code class="literal">digit</code>,
<code class="literal">graph</code>, <code class="literal">lower</code>,
<code class="literal">print</code>, <code class="literal">punct</code>,
<code class="literal">space</code>, <code class="literal">upper</code>,
<code class="literal">xdigit</code>. These stand for the character classes
defined in
<span class="citerefentry"><span class="refentrytitle">ctype</span></span>.
A locale can provide others. A character class cannot be used as
an endpoint of a range.
</p><p>
There are two special cases of bracket expressions: the bracket
expressions <code class="literal">[[:<:]]</code> and
<code class="literal">[[:>:]]</code> are constraints,
matching empty strings at the beginning
and end of a word respectively. A word is defined as a sequence
of word characters that is neither preceded nor followed by word
characters. A word character is an <code class="literal">alnum</code> character (as
defined by
<span class="citerefentry"><span class="refentrytitle">ctype</span></span>)
or an underscore. This is an extension, compatible with but not
specified by <acronym class="acronym">POSIX</acronym> 1003.2, and should be used with
caution in software intended to be portable to other systems.
The constraint escapes described below are usually preferable; they
are no more standard, but are easier to type.
</p></div><div class="sect3" id="POSIX-ESCAPE-SEQUENCES"><div class="titlepage"><div><div><h4 class="title">9.7.3.3. Regular Expression Escapes</h4></div></div></div><p>
<em class="firstterm">Escapes</em> are special sequences beginning with <code class="literal">\</code>
followed by an alphanumeric character. Escapes come in several varieties:
character entry, class shorthands, constraint escapes, and back references.
A <code class="literal">\</code> followed by an alphanumeric character but not constituting
a valid escape is illegal in AREs.
In EREs, there are no escapes: outside a bracket expression,
a <code class="literal">\</code> followed by an alphanumeric character merely stands for
that character as an ordinary character, and inside a bracket expression,
<code class="literal">\</code> is an ordinary character.
(The latter is the one actual incompatibility between EREs and AREs.)
</p><p>
<em class="firstterm">Character-entry escapes</em> exist to make it easier to specify
non-printing and other inconvenient characters in REs. They are
shown in <a class="xref" href="functions-matching.html#POSIX-CHARACTER-ENTRY-ESCAPES-TABLE" title="Table 9.18. Regular Expression Character-entry Escapes">Table 9.18</a>.
</p><p>
<em class="firstterm">Class-shorthand escapes</em> provide shorthands for certain
commonly-used character classes. They are
shown in <a class="xref" href="functions-matching.html#POSIX-CLASS-SHORTHAND-ESCAPES-TABLE" title="Table 9.19. Regular Expression Class-shorthand Escapes">Table 9.19</a>.
</p><p>
A <em class="firstterm">constraint escape</em> is a constraint,
matching the empty string if specific conditions are met,
written as an escape. They are
shown in <a class="xref" href="functions-matching.html#POSIX-CONSTRAINT-ESCAPES-TABLE" title="Table 9.20. Regular Expression Constraint Escapes">Table 9.20</a>.
</p><p>
A <em class="firstterm">back reference</em> (<code class="literal">\</code><em class="replaceable"><code>n</code></em>) matches the
same string matched by the previous parenthesized subexpression specified
by the number <em class="replaceable"><code>n</code></em>
(see <a class="xref" href="functions-matching.html#POSIX-CONSTRAINT-BACKREF-TABLE" title="Table 9.21. Regular Expression Back References">Table 9.21</a>). For example,
<code class="literal">([bc])\1</code> matches <code class="literal">bb</code> or <code class="literal">cc</code>
but not <code class="literal">bc</code> or <code class="literal">cb</code>.
The subexpression must entirely precede the back reference in the RE.
Subexpressions are numbered in the order of their leading parentheses.
Non-capturing parentheses do not define subexpressions.
</p><div class="table" id="POSIX-CHARACTER-ENTRY-ESCAPES-TABLE"><p class="title"><strong>Table 9.18. Regular Expression Character-entry Escapes</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Character-entry Escapes" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Escape</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">\a</code> </td><td> alert (bell) character, as in C </td></tr><tr><td> <code class="literal">\b</code> </td><td> backspace, as in C </td></tr><tr><td> <code class="literal">\B</code> </td><td> synonym for backslash (<code class="literal">\</code>) to help reduce the need for backslash
doubling </td></tr><tr><td> <code class="literal">\c</code><em class="replaceable"><code>X</code></em> </td><td> (where <em class="replaceable"><code>X</code></em> is any character) the character whose
low-order 5 bits are the same as those of
<em class="replaceable"><code>X</code></em>, and whose other bits are all zero </td></tr><tr><td> <code class="literal">\e</code> </td><td> the character whose collating-sequence name
is <code class="literal">ESC</code>,
or failing that, the character with octal value <code class="literal">033</code> </td></tr><tr><td> <code class="literal">\f</code> </td><td> form feed, as in C </td></tr><tr><td> <code class="literal">\n</code> </td><td> newline, as in C </td></tr><tr><td> <code class="literal">\r</code> </td><td> carriage return, as in C </td></tr><tr><td> <code class="literal">\t</code> </td><td> horizontal tab, as in C </td></tr><tr><td> <code class="literal">\u</code><em class="replaceable"><code>wxyz</code></em> </td><td> (where <em class="replaceable"><code>wxyz</code></em> is exactly four hexadecimal digits)
the character whose hexadecimal value is
<code class="literal">0x</code><em class="replaceable"><code>wxyz</code></em>
</td></tr><tr><td> <code class="literal">\U</code><em class="replaceable"><code>stuvwxyz</code></em> </td><td> (where <em class="replaceable"><code>stuvwxyz</code></em> is exactly eight hexadecimal
digits)
the character whose hexadecimal value is
<code class="literal">0x</code><em class="replaceable"><code>stuvwxyz</code></em>
</td></tr><tr><td> <code class="literal">\v</code> </td><td> vertical tab, as in C </td></tr><tr><td> <code class="literal">\x</code><em class="replaceable"><code>hhh</code></em> </td><td> (where <em class="replaceable"><code>hhh</code></em> is any sequence of hexadecimal
digits)
the character whose hexadecimal value is
<code class="literal">0x</code><em class="replaceable"><code>hhh</code></em>
(a single character no matter how many hexadecimal digits are used)
</td></tr><tr><td> <code class="literal">\0</code> </td><td> the character whose value is <code class="literal">0</code> (the null byte)</td></tr><tr><td> <code class="literal">\</code><em class="replaceable"><code>xy</code></em> </td><td> (where <em class="replaceable"><code>xy</code></em> is exactly two octal digits,
and is not a <em class="firstterm">back reference</em>)
the character whose octal value is
<code class="literal">0</code><em class="replaceable"><code>xy</code></em> </td></tr><tr><td> <code class="literal">\</code><em class="replaceable"><code>xyz</code></em> </td><td> (where <em class="replaceable"><code>xyz</code></em> is exactly three octal digits,
and is not a <em class="firstterm">back reference</em>)
the character whose octal value is
<code class="literal">0</code><em class="replaceable"><code>xyz</code></em> </td></tr></tbody></table></div></div><br class="table-break" /><p>
Hexadecimal digits are <code class="literal">0</code>-<code class="literal">9</code>,
<code class="literal">a</code>-<code class="literal">f</code>, and <code class="literal">A</code>-<code class="literal">F</code>.
Octal digits are <code class="literal">0</code>-<code class="literal">7</code>.
</p><p>
Numeric character-entry escapes specifying values outside the ASCII range
(0-127) have meanings dependent on the database encoding. When the
encoding is UTF-8, escape values are equivalent to Unicode code points,
for example <code class="literal">\u1234</code> means the character <code class="literal">U+1234</code>.
For other multibyte encodings, character-entry escapes usually just
specify the concatenation of the byte values for the character. If the
escape value does not correspond to any legal character in the database
encoding, no error will be raised, but it will never match any data.
</p><p>
The character-entry escapes are always taken as ordinary characters.
For example, <code class="literal">\135</code> is <code class="literal">]</code> in ASCII, but
<code class="literal">\135</code> does not terminate a bracket expression.
</p><div class="table" id="POSIX-CLASS-SHORTHAND-ESCAPES-TABLE"><p class="title"><strong>Table 9.19. Regular Expression Class-shorthand Escapes</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Class-shorthand Escapes" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Escape</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">\d</code> </td><td> <code class="literal">[[:digit:]]</code> </td></tr><tr><td> <code class="literal">\s</code> </td><td> <code class="literal">[[:space:]]</code> </td></tr><tr><td> <code class="literal">\w</code> </td><td> <code class="literal">[[:alnum:]_]</code>
(note underscore is included) </td></tr><tr><td> <code class="literal">\D</code> </td><td> <code class="literal">[^[:digit:]]</code> </td></tr><tr><td> <code class="literal">\S</code> </td><td> <code class="literal">[^[:space:]]</code> </td></tr><tr><td> <code class="literal">\W</code> </td><td> <code class="literal">[^[:alnum:]_]</code>
(note underscore is included) </td></tr></tbody></table></div></div><br class="table-break" /><p>
Within bracket expressions, <code class="literal">\d</code>, <code class="literal">\s</code>,
and <code class="literal">\w</code> lose their outer brackets,
and <code class="literal">\D</code>, <code class="literal">\S</code>, and <code class="literal">\W</code> are illegal.
(So, for example, <code class="literal">[a-c\d]</code> is equivalent to
<code class="literal">[a-c[:digit:]]</code>.
Also, <code class="literal">[a-c\D]</code>, which is equivalent to
<code class="literal">[a-c^[:digit:]]</code>, is illegal.)
</p><div class="table" id="POSIX-CONSTRAINT-ESCAPES-TABLE"><p class="title"><strong>Table 9.20. Regular Expression Constraint Escapes</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Constraint Escapes" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Escape</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">\A</code> </td><td> matches only at the beginning of the string
(see <a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a> for how this differs from
<code class="literal">^</code>) </td></tr><tr><td> <code class="literal">\m</code> </td><td> matches only at the beginning of a word </td></tr><tr><td> <code class="literal">\M</code> </td><td> matches only at the end of a word </td></tr><tr><td> <code class="literal">\y</code> </td><td> matches only at the beginning or end of a word </td></tr><tr><td> <code class="literal">\Y</code> </td><td> matches only at a point that is not the beginning or end of a
word </td></tr><tr><td> <code class="literal">\Z</code> </td><td> matches only at the end of the string
(see <a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a> for how this differs from
<code class="literal">$</code>) </td></tr></tbody></table></div></div><br class="table-break" /><p>
A word is defined as in the specification of
<code class="literal">[[:<:]]</code> and <code class="literal">[[:>:]]</code> above.
Constraint escapes are illegal within bracket expressions.
</p><div class="table" id="POSIX-CONSTRAINT-BACKREF-TABLE"><p class="title"><strong>Table 9.21. Regular Expression Back References</strong></p><div class="table-contents"><table class="table" summary="Regular Expression Back References" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Escape</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">\</code><em class="replaceable"><code>m</code></em> </td><td> (where <em class="replaceable"><code>m</code></em> is a nonzero digit)
a back reference to the <em class="replaceable"><code>m</code></em>'th subexpression </td></tr><tr><td> <code class="literal">\</code><em class="replaceable"><code>mnn</code></em> </td><td> (where <em class="replaceable"><code>m</code></em> is a nonzero digit, and
<em class="replaceable"><code>nn</code></em> is some more digits, and the decimal value
<em class="replaceable"><code>mnn</code></em> is not greater than the number of closing capturing
parentheses seen so far)
a back reference to the <em class="replaceable"><code>mnn</code></em>'th subexpression </td></tr></tbody></table></div></div><br class="table-break" /><div class="note"><h3 class="title">Note</h3><p>
There is an inherent ambiguity between octal character-entry
escapes and back references, which is resolved by the following heuristics,
as hinted at above.
A leading zero always indicates an octal escape.
A single non-zero digit, not followed by another digit,
is always taken as a back reference.
A multi-digit sequence not starting with a zero is taken as a back
reference if it comes after a suitable subexpression
(i.e., the number is in the legal range for a back reference),
and otherwise is taken as octal.
</p></div></div><div class="sect3" id="POSIX-METASYNTAX"><div class="titlepage"><div><div><h4 class="title">9.7.3.4. Regular Expression Metasyntax</h4></div></div></div><p>
In addition to the main syntax described above, there are some special
forms and miscellaneous syntactic facilities available.
</p><p>
An RE can begin with one of two special <em class="firstterm">director</em> prefixes.
If an RE begins with <code class="literal">***:</code>,
the rest of the RE is taken as an ARE. (This normally has no effect in
<span class="productname">PostgreSQL</span>, since REs are assumed to be AREs;
but it does have an effect if ERE or BRE mode had been specified by
the <em class="replaceable"><code>flags</code></em> parameter to a regex function.)
If an RE begins with <code class="literal">***=</code>,
the rest of the RE is taken to be a literal string,
with all characters considered ordinary characters.
</p><p>
An ARE can begin with <em class="firstterm">embedded options</em>:
a sequence <code class="literal">(?</code><em class="replaceable"><code>xyz</code></em><code class="literal">)</code>
(where <em class="replaceable"><code>xyz</code></em> is one or more alphabetic characters)
specifies options affecting the rest of the RE.
These options override any previously determined options —
in particular, they can override the case-sensitivity behavior implied by
a regex operator, or the <em class="replaceable"><code>flags</code></em> parameter to a regex
function.
The available option letters are
shown in <a class="xref" href="functions-matching.html#POSIX-EMBEDDED-OPTIONS-TABLE" title="Table 9.22. ARE Embedded-option Letters">Table 9.22</a>.
Note that these same option letters are used in the <em class="replaceable"><code>flags</code></em>
parameters of regex functions.
</p><div class="table" id="POSIX-EMBEDDED-OPTIONS-TABLE"><p class="title"><strong>Table 9.22. ARE Embedded-option Letters</strong></p><div class="table-contents"><table class="table" summary="ARE Embedded-option Letters" border="1"><colgroup><col /><col /></colgroup><thead><tr><th>Option</th><th>Description</th></tr></thead><tbody><tr><td> <code class="literal">b</code> </td><td> rest of RE is a BRE </td></tr><tr><td> <code class="literal">c</code> </td><td> case-sensitive matching (overrides operator type) </td></tr><tr><td> <code class="literal">e</code> </td><td> rest of RE is an ERE </td></tr><tr><td> <code class="literal">i</code> </td><td> case-insensitive matching (see
<a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a>) (overrides operator type) </td></tr><tr><td> <code class="literal">m</code> </td><td> historical synonym for <code class="literal">n</code> </td></tr><tr><td> <code class="literal">n</code> </td><td> newline-sensitive matching (see
<a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a>) </td></tr><tr><td> <code class="literal">p</code> </td><td> partial newline-sensitive matching (see
<a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a>) </td></tr><tr><td> <code class="literal">q</code> </td><td> rest of RE is a literal (<span class="quote">“<span class="quote">quoted</span>”</span>) string, all ordinary
characters </td></tr><tr><td> <code class="literal">s</code> </td><td> non-newline-sensitive matching (default) </td></tr><tr><td> <code class="literal">t</code> </td><td> tight syntax (default; see below) </td></tr><tr><td> <code class="literal">w</code> </td><td> inverse partial newline-sensitive (<span class="quote">“<span class="quote">weird</span>”</span>) matching
(see <a class="xref" href="functions-matching.html#POSIX-MATCHING-RULES" title="9.7.3.5. Regular Expression Matching Rules">Section 9.7.3.5</a>) </td></tr><tr><td> <code class="literal">x</code> </td><td> expanded syntax (see below) </td></tr></tbody></table></div></div><br class="table-break" /><p>
Embedded options take effect at the <code class="literal">)</code> terminating the sequence.
They can appear only at the start of an ARE (after the
<code class="literal">***:</code> director if any).
</p><p>
In addition to the usual (<em class="firstterm">tight</em>) RE syntax, in which all
characters are significant, there is an <em class="firstterm">expanded</em> syntax,
available by specifying the embedded <code class="literal">x</code> option.
In the expanded syntax,
white-space characters in the RE are ignored, as are
all characters between a <code class="literal">#</code>
and the following newline (or the end of the RE). This
permits paragraphing and commenting a complex RE.
There are three exceptions to that basic rule:
</p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
a white-space character or <code class="literal">#</code> preceded by <code class="literal">\</code> is
retained
</p></li><li class="listitem"><p>
white space or <code class="literal">#</code> within a bracket expression is retained
</p></li><li class="listitem"><p>
white space and comments cannot appear within multi-character symbols,
such as <code class="literal">(?:</code>
</p></li></ul></div><p>
For this purpose, white-space characters are blank, tab, newline, and
any character that belongs to the <em class="replaceable"><code>space</code></em> character class.
</p><p>
Finally, in an ARE, outside bracket expressions, the sequence
<code class="literal">(?#</code><em class="replaceable"><code>ttt</code></em><code class="literal">)</code>
(where <em class="replaceable"><code>ttt</code></em> is any text not containing a <code class="literal">)</code>)
is a comment, completely ignored.
Again, this is not allowed between the characters of
multi-character symbols, like <code class="literal">(?:</code>.
Such comments are more a historical artifact than a useful facility,
and their use is deprecated; use the expanded syntax instead.
</p><p>
<span class="emphasis"><em>None</em></span> of these metasyntax extensions is available if
an initial <code class="literal">***=</code> director
has specified that the user's input be treated as a literal string
rather than as an RE.
</p></div><div class="sect3" id="POSIX-MATCHING-RULES"><div class="titlepage"><div><div><h4 class="title">9.7.3.5. Regular Expression Matching Rules</h4></div></div></div><p>
In the event that an RE could match more than one substring of a given
string, the RE matches the one starting earliest in the string.
If the RE could match more than one substring starting at that point,
either the longest possible match or the shortest possible match will
be taken, depending on whether the RE is <em class="firstterm">greedy</em> or
<em class="firstterm">non-greedy</em>.
</p><p>
Whether an RE is greedy or not is determined by the following rules:
</p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
Most atoms, and all constraints, have no greediness attribute (because
they cannot match variable amounts of text anyway).
</p></li><li class="listitem"><p>
Adding parentheses around an RE does not change its greediness.
</p></li><li class="listitem"><p>
A quantified atom with a fixed-repetition quantifier
(<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">}</code>
or
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">}?</code>)
has the same greediness (possibly none) as the atom itself.
</p></li><li class="listitem"><p>
A quantified atom with other normal quantifiers (including
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,</code><em class="replaceable"><code>n</code></em><code class="literal">}</code>
with <em class="replaceable"><code>m</code></em> equal to <em class="replaceable"><code>n</code></em>)
is greedy (prefers longest match).
</p></li><li class="listitem"><p>
A quantified atom with a non-greedy quantifier (including
<code class="literal">{</code><em class="replaceable"><code>m</code></em><code class="literal">,</code><em class="replaceable"><code>n</code></em><code class="literal">}?</code>
with <em class="replaceable"><code>m</code></em> equal to <em class="replaceable"><code>n</code></em>)
is non-greedy (prefers shortest match).
</p></li><li class="listitem"><p>
A branch — that is, an RE that has no top-level
<code class="literal">|</code> operator — has the same greediness as the first
quantified atom in it that has a greediness attribute.
</p></li><li class="listitem"><p>
An RE consisting of two or more branches connected by the
<code class="literal">|</code> operator is always greedy.
</p></li></ul></div><p>
</p><p>
The above rules associate greediness attributes not only with individual
quantified atoms, but with branches and entire REs that contain quantified
atoms. What that means is that the matching is done in such a way that
the branch, or whole RE, matches the longest or shortest possible
substring <span class="emphasis"><em>as a whole</em></span>. Once the length of the entire match
is determined, the part of it that matches any particular subexpression
is determined on the basis of the greediness attribute of that
subexpression, with subexpressions starting earlier in the RE taking
priority over ones starting later.
</p><p>
An example of what this means:
</p><pre class="screen">
SELECT SUBSTRING('XY1234Z', 'Y*([0-9]{1,3})');
<em class="lineannotation"><span class="lineannotation">Result: </span></em><code class="computeroutput">123</code>
SELECT SUBSTRING('XY1234Z', 'Y*?([0-9]{1,3})');
<em class="lineannotation"><span class="lineannotation">Result: </span></em><code class="computeroutput">1</code>
</pre><p>
In the first case, the RE as a whole is greedy because <code class="literal">Y*</code>
is greedy. It can match beginning at the <code class="literal">Y</code>, and it matches
the longest possible string starting there, i.e., <code class="literal">Y123</code>.
The output is the parenthesized part of that, or <code class="literal">123</code>.
In the second case, the RE as a whole is non-greedy because <code class="literal">Y*?</code>
is non-greedy. It can match beginning at the <code class="literal">Y</code>, and it matches
the shortest possible string starting there, i.e., <code class="literal">Y1</code>.
The subexpression <code class="literal">[0-9]{1,3}</code> is greedy but it cannot change
the decision as to the overall match length; so it is forced to match
just <code class="literal">1</code>.
</p><p>
In short, when an RE contains both greedy and non-greedy subexpressions,
the total match length is either as long as possible or as short as
possible, according to the attribute assigned to the whole RE. The
attributes assigned to the subexpressions only affect how much of that
match they are allowed to <span class="quote">“<span class="quote">eat</span>”</span> relative to each other.
</p><p>
The quantifiers <code class="literal">{1,1}</code> and <code class="literal">{1,1}?</code>
can be used to force greediness or non-greediness, respectively,
on a subexpression or a whole RE.
This is useful when you need the whole RE to have a greediness attribute
different from what's deduced from its elements. As an example,
suppose that we are trying to separate a string containing some digits
into the digits and the parts before and after them. We might try to
do that like this:
</p><pre class="screen">
SELECT regexp_match('abc01234xyz', '(.*)(\d+)(.*)');
<em class="lineannotation"><span class="lineannotation">Result: </span></em><code class="computeroutput">{abc0123,4,xyz}</code>
</pre><p>
That didn't work: the first <code class="literal">.*</code> is greedy so
it <span class="quote">“<span class="quote">eats</span>”</span> as much as it can, leaving the <code class="literal">\d+</code> to
match at the last possible place, the last digit. We might try to fix
that by making it non-greedy:
</p><pre class="screen">
SELECT regexp_match('abc01234xyz', '(.*?)(\d+)(.*)');
<em class="lineannotation"><span class="lineannotation">Result: </span></em><code class="computeroutput">{abc,0,""}</code>
</pre><p>
That didn't work either, because now the RE as a whole is non-greedy
and so it ends the overall match as soon as possible. We can get what
we want by forcing the RE as a whole to be greedy:
</p><pre class="screen">
SELECT regexp_match('abc01234xyz', '(?:(.*?)(\d+)(.*)){1,1}');
<em class="lineannotation"><span class="lineannotation">Result: </span></em><code class="computeroutput">{abc,01234,xyz}</code>
</pre><p>
Controlling the RE's overall greediness separately from its components'
greediness allows great flexibility in handling variable-length patterns.
</p><p>
When deciding what is a longer or shorter match,
match lengths are measured in characters, not collating elements.
An empty string is considered longer than no match at all.
For example:
<code class="literal">bb*</code>
matches the three middle characters of <code class="literal">abbbc</code>;
<code class="literal">(week|wee)(night|knights)</code>
matches all ten characters of <code class="literal">weeknights</code>;
when <code class="literal">(.*).*</code>
is matched against <code class="literal">abc</code> the parenthesized subexpression
matches all three characters; and when
<code class="literal">(a*)*</code> is matched against <code class="literal">bc</code>
both the whole RE and the parenthesized
subexpression match an empty string.
</p><p>
If case-independent matching is specified,
the effect is much as if all case distinctions had vanished from the
alphabet.
When an alphabetic that exists in multiple cases appears as an
ordinary character outside a bracket expression, it is effectively
transformed into a bracket expression containing both cases,
e.g., <code class="literal">x</code> becomes <code class="literal">[xX]</code>.
When it appears inside a bracket expression, all case counterparts
of it are added to the bracket expression, e.g.,
<code class="literal">[x]</code> becomes <code class="literal">[xX]</code>
and <code class="literal">[^x]</code> becomes <code class="literal">[^xX]</code>.
</p><p>
If newline-sensitive matching is specified, <code class="literal">.</code>
and bracket expressions using <code class="literal">^</code>
will never match the newline character
(so that matches will never cross newlines unless the RE
explicitly arranges it)
and <code class="literal">^</code> and <code class="literal">$</code>
will match the empty string after and before a newline
respectively, in addition to matching at beginning and end of string
respectively.
But the ARE escapes <code class="literal">\A</code> and <code class="literal">\Z</code>
continue to match beginning or end of string <span class="emphasis"><em>only</em></span>.
</p><p>
If partial newline-sensitive matching is specified,
this affects <code class="literal">.</code> and bracket expressions
as with newline-sensitive matching, but not <code class="literal">^</code>
and <code class="literal">$</code>.
</p><p>
If inverse partial newline-sensitive matching is specified,
this affects <code class="literal">^</code> and <code class="literal">$</code>
as with newline-sensitive matching, but not <code class="literal">.</code>
and bracket expressions.
This isn't very useful but is provided for symmetry.
</p></div><div class="sect3" id="POSIX-LIMITS-COMPATIBILITY"><div class="titlepage"><div><div><h4 class="title">9.7.3.6. Limits and Compatibility</h4></div></div></div><p>
No particular limit is imposed on the length of REs in this
implementation. However,
programs intended to be highly portable should not employ REs longer
than 256 bytes,
as a POSIX-compliant implementation can refuse to accept such REs.
</p><p>
The only feature of AREs that is actually incompatible with
POSIX EREs is that <code class="literal">\</code> does not lose its special
significance inside bracket expressions.
All other ARE features use syntax which is illegal or has
undefined or unspecified effects in POSIX EREs;
the <code class="literal">***</code> syntax of directors likewise is outside the POSIX
syntax for both BREs and EREs.
</p><p>
Many of the ARE extensions are borrowed from Perl, but some have
been changed to clean them up, and a few Perl extensions are not present.
Incompatibilities of note include <code class="literal">\b</code>, <code class="literal">\B</code>,
the lack of special treatment for a trailing newline,
the addition of complemented bracket expressions to the things
affected by newline-sensitive matching,
the restrictions on parentheses and back references in lookahead/lookbehind
constraints, and the longest/shortest-match (rather than first-match)
matching semantics.
</p><p>
Two significant incompatibilities exist between AREs and the ERE syntax
recognized by pre-7.4 releases of <span class="productname">PostgreSQL</span>:
</p><div class="itemizedlist"><ul class="itemizedlist" style="list-style-type: disc; "><li class="listitem"><p>
In AREs, <code class="literal">\</code> followed by an alphanumeric character is either
an escape or an error, while in previous releases, it was just another
way of writing the alphanumeric.
This should not be much of a problem because there was no reason to
write such a sequence in earlier releases.
</p></li><li class="listitem"><p>
In AREs, <code class="literal">\</code> remains a special character within
<code class="literal">[]</code>, so a literal <code class="literal">\</code> within a bracket
expression must be written <code class="literal">\\</code>.
</p></li></ul></div><p>
</p></div><div class="sect3" id="POSIX-BASIC-REGEXES"><div class="titlepage"><div><div><h4 class="title">9.7.3.7. Basic Regular Expressions</h4></div></div></div><p>
BREs differ from EREs in several respects.
In BREs, <code class="literal">|</code>, <code class="literal">+</code>, and <code class="literal">?</code>
are ordinary characters and there is no equivalent
for their functionality.
The delimiters for bounds are
<code class="literal">\{</code> and <code class="literal">\}</code>,
with <code class="literal">{</code> and <code class="literal">}</code>
by themselves ordinary characters.
The parentheses for nested subexpressions are
<code class="literal">\(</code> and <code class="literal">\)</code>,
with <code class="literal">(</code> and <code class="literal">)</code> by themselves ordinary characters.
<code class="literal">^</code> is an ordinary character except at the beginning of the
RE or the beginning of a parenthesized subexpression,
<code class="literal">$</code> is an ordinary character except at the end of the
RE or the end of a parenthesized subexpression,
and <code class="literal">*</code> is an ordinary character if it appears at the beginning
of the RE or the beginning of a parenthesized subexpression
(after a possible leading <code class="literal">^</code>).
Finally, single-digit back references are available, and
<code class="literal">\<</code> and <code class="literal">\></code>
are synonyms for
<code class="literal">[[:<:]]</code> and <code class="literal">[[:>:]]</code>
respectively; no other escapes are available in BREs.
</p></div></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="functions-bitstring.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="functions.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="functions-formatting.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">9.6. Bit String Functions and Operators </td><td width="20%" align="center"><a accesskey="h" href="index.html">Home</a></td><td width="40%" align="right" valign="top"> 9.8. Data Type Formatting Functions</td></tr></table></div></body></html>
