multi_delimiter

The multi_delimiter template splits the input wherever it encounters any of several delimiter characters, given as the DELIMITERS list. It suits fields that mix separators — for example splitting key:value; key2:value2 on :, ; and space yields the individual keys and values.

Apart from accepting a set of delimiters rather than one, it behaves like delimiter: it emits the pieces verbatim, so chain it into a pipeline if you also need case folding or stemming.

Options

Option	Type	Default	Description
DELIMITERS	string list	required	Comma-separated delimiters (e.g., '":", ";", " "')

Tokenization

The template cuts the input wherever it finds any character in the DELIMITERS set and emits the pieces verbatim — no further analysis. It behaves like delimiter but accepts several separators at once, which suits fields that mix them.

Input	Delimiters	Tokens
key:value; key2:value2	: ; space	{key,value,key2,value2}
2026-06-18 logs/app	/ - space	{2026,06,18,logs,app}

Splitting key:value; key2:value2 on the colon, semicolon and space separators recovers the individual keys and values in one pass. Preview the split with ts_lexize:

Query

CREATE TEXT SEARCH DICTIONARY tok_multi_delim (    template = 'multi_delimiter',    delimiters = '":", ";", " "');
SELECT ts_lexize('tok_multi_delim', 'key:value; key2:value2');

Result

 ts_lexize------------------------- {key,value,key2,value2}

Examples

Query

CREATE TEXT SEARCH DICTIONARY multi_delim (    template = 'multi_delimiter',    delimiters = '":", ";", " ", ".", "|"');

Because the delimiter list is itself comma-separated, a comma cannot be used as a delimiter here. To split on a single comma, use the delimiter template instead.

See also

• delimiter — split on a single delimiter
• CREATE TEXT SEARCH DICTIONARY