$smwgFulltextSearchMinTokenSize

Configuration parameter details:
Name	$smwgFulltextSearchMinTokenSize
Description	Sets the minimum word/token length to help to decide whether MATCH or LIKE operators are to be used for a condition statement
Default setting	3
Software	Semantic MediaWiki
Since version	2.5.0
Until version	still available
Configuration	Full-text search · Experimental
Keyword	full-text search · data store · relational database · sql store · sql database · experimental

$smwgFulltextSearchMinTokenSize is a configuration parameter that sets the minimum word/token length to help to decide whether MATCH or LIKE operators are to be used for a condition statement of the full-text search. This helps to switch back to LIKE in cases where the minimum threshold is not applicable. The configuration parameter is specific to the relational data stores MySQL and MariaDB. The higher the number set for this configuration parameter the faster the queries will be. However the default setting appears to be the most pragmatic in most cases when querying for meaningful content. The configuration parameter was introduced in Semantic MediaWiki 2.5.0Released on 14 March 2017 and compatible with MW 1.23.0 - 1.29.x..1

This configuration parameter only takes effect if full-text search was enabled via configuration parameter $smwgEnabledFulltextSearchSets whether full-text search support for properties may be used.

Default setting

$smwgFulltextSearchMinTokenSize = 3;

This means that the minimum length of words stored in the InnoDB FULLTEXT index is "3".

Changing the default setting

If this configuration parameter is changed it must be set to an integer between "1" and "16" corresponding to either MySQL's configuration parameter innodb_ft_min_token_size (integer between "0" and "16") or configuration parameter ft_min_word_len (integer higher than "1").

Maintenance script "rebuildFulltextSearchTable.php"Allows to rebuild the full text search data table must be run after changing the setting of this configuration parameter.

To modify the setting to this configuration parameter, add one of the following lines to your "LocalSettings.php" file after the enableSemantics() call:

Increase the the minimum length of words stored

$smwgFulltextSearchMinTokenSize = 5;

This means that the minimum length of words stored in the InnoDB FULLTEXT index is "5". This reduces the size of the index, thus speeding up queries, by omitting common words that are unlikely to be significant in a search context, such as the English words "a", "to" and "and".

Reduce the the minimum length of words stored

$smwgFulltextSearchMinTokenSize = 1;

This means that the minimum length of words stored in the InnoDB FULLTEXT index is "1". Only recommended for a CJK languages (Chinese, Japanese, Korean).

See also

General information

• Help page on full-text search
• Help page on how to use full-text search

Related configuration parameters

• Help page on configuration parameter $smwgEnabledFulltextSearchSets whether full-text search support for properties may be used − Allows to enable the full-text search feature
• Help page on configuration parameter $smwgFulltextDeferredUpdateSets the number of expected full-text search index updates − Allows to throttle the number of expected index updates
• Help page on configuration parameter $smwgFulltextSearchTableOptionsSets the full-text search table options to use during installation or update − Allows to set database related options
• Help page on configuration parameter $smwgFulltextLanguageDetectionSets which languages to detect for the full-text search from an indexable text − Allows to detect a language (experimental setting)
• Help page on configuration parameter $smwgFulltextSearchIndexableDataTypesSets which datatypes are allowed to be indexed using the full-text search − Allows to list datatypes that should be indexed
• Help page on configuration parameter $smwgFulltextSearchPropertyExemptionListSets the property keys for which value assignments are being exempted from the full-text indexing − Allows to list properties that should be not be indexed
• Help page on option SMW_FIELDT_CHAR_LONG to configuration parameter $smwgDVFeatures − Allows to extend the searchable field length to 300 characters without the need to enable the