Full text search

The relation function extracts field values and content and allows to get insight about the data that is stored in the wiki.

When the user browses on the wiki, however, he does not see the data, but the rendered page in HTML. When the user searches with the search box, then not the original data is relevant, but the rendered page.

Full text search goes this new approach, using a technology FTS3 from SQLite. Besides being more relevant, the search is also very fast.

Full text search finds all pages that match the word you search for. You can use binary logic

"wiki" finds all pages that have words that contain "wiki" ("Wiki", "wikilanguage", "Sofawiki")
"paris berlin" finds all pages that contain both "paris" and "berlin"
"paris|london" finds all pages that contain either "paris" or "london"

Note that SofaWiki uses to modifications of the syntax of FTS3

There are now keywords AND and OR. AND is implicite by space, OR is expressed by the pipe character
There is no "*" extension. SofaWiki captures all pages that contain the word, the word does not have to match exactly.
Match exactly and NEAR are not supported.

You can use full text search, if the PHP version you use supports FTS3 (most do) and you set $swUseFulltext in configuration.php to true.

Please note that the index gets some time to build up. A page is only included in the index if it is viewed.
Only pages with the namespaces defined in $swSearchNamespaces are included.
Pages with  are not included.

For technical details see fulltext.php