Full-text search in SQLite

Thanks to Comrade Scott Hess of Google, SQLite also launched the full-text search function.

SQLite creates a virtual table to split the input text into B trees based on rules such as spaces and punctuation. This improves the search speed.

How to use this function and how to use it in terms of performance can be further studied.

 

Here I will report on my research results:

1. The full-text search of SQLite is roughly divided into words by space. IfArticleIf the word "ASDF" exists, you will not want to use "SD" to find the content, because it is organized by words as a unit;

 

2. Since it is space-based word segmentation, Chinese characters are naturally not supported, unless you separate each word in an article with spaces or punctuation marks;

 

3. SQLite provides a customizable lexical analyzer. For detailed usage, see the document;

 

4. for more information about Porter, see the source code file "fts3_porter.c". Multiple word shapes of a word are unified into one basic word according to the Basic English specifications, with semantic support.

 

To sum up, we have not provided true semantic word segmentation, so it is up to you to make great efforts.

Link to the official website for full-text search:Http://www.sqlite.org/fts3.html