Full Text for SQLite3 0.2.2 review

by rbytes.net on

Full Text for SQLite3 is a full text indexer for datas stored into a sqlite3 database

License: BSD License
File size: 0K
Developer: Pierre Aubert
0 stars award from rbytes.net

Full Text for SQLite3 is a full text indexer for datas stored into a sqlite3 database. Full text means that the text fields stored into the database can be split word by word and stored in a form suitable for lookup.

You can then create easy SQL queries for searching for cells from yours indexed tables containing the words from the query.

More details: words handling

You have of course remark that the matching as been done on unfolded letters, that is word 'New' match word 'New' and word 'new'. This is a classical design which works well for english, but you can decide otherwise. Similary the city name 'New-York' has been split in two. You may agree or not.

In language with diacritics, you have to decide how you want the matching done:
In french, word 'cuvre' with the ligature oe may match 'oeuvre' for example. In german a '?' may match 'ss'.

FT3 provides an API for defining your own matching. This is also dependant of the underlying encoding. Currently we only support UTF-8 which is used by default inside sqlite3.

More details: stemmed words

A stem helps in providing a match for single/plural words or conjugated forms. It is expected that the stemmed form of a word will be the same for a noun plural or singular and that most forms of a verb will give the same stem. The algorithm which computes the stem is of course language dependant. In english, it use the well known Porter algorithm, and variants of it in other indo-europeen languages. The language supported are those of the library in contrib/stemmed directory.

For example in english, words 'cars' and 'car' will match. For exemple, in french words 'saisons' and 'saison' will have the same stem.
The algorithm is not perfect in particular for language with a lot of exceptions like french.

More details: similar words

In some cases, it is useful to have words related by one way or one other. Similar words can be words with the same phonetic for example. That is usefull with words like 'schwartzenegger' or 'jonasz'.

What's New in This Release:
Version 0.2.1 shipped without an include file (sqlite3xx.hpp). It has been re-added.

Full Text for SQLite3 0.2.2 search tags