PostgreSQL 10.0 preview 功能增强 - JSON 内容全文检索

标签

PostgreSQL , 10.0 , jsonb , json , 全文检索

背景

PostgreSQL 10.0 支持JSON内容的全文检索了。

同样支持ts rank和phrase 索引哦。

ts rank, phrase请参考

《从难缠的模糊查询聊开 - PostgreSQL独门绝招之一 GIN , GiST , SP-GiST , RUM 索引原理与技术背景》

Hi all  

I would like to propose patch with a set of new small functions for fts in
case of
jsonb data type:  

* to_tsvector(config, jsonb) - make a tsvector from all string values and
  elements of jsonb object. To prevent the situation, when tsquery can find
  a
  phrase consisting of lexemes from two different values/elements, this
  function will add an increment to position of each lexeme from every new
  value/element.  

* ts_headline(config, jsonb, tsquery, options) - generate a headline
directly
  from jsonb object  

Here are the examples how they work:  

=# select to_tsvector('{"a": "aaa bbb", "b": ["ccc ddd"], "c": {"d": "eee
fff"}}'::jsonb);
                   to_tsvector
-------------------------------------------------
 'aaa':1 'bbb':2 'ccc':4 'ddd':5 'eee':7 'fff':8
(1 row)  

=# select ts_headline('english', '{"a": "aaa bbb", "b": {"c": "ccc
ddd"}}'::jsonb, tsquery('bbb & ddd & hhh'), 'StartSel = <, StopSel = >');
     ts_headline
----------------------
 aaa <bbb> ccc <ddd>
(1 row)  

Any comments or suggestions?

这个patch的讨论，详见邮件组，本文末尾URL。

PostgreSQL社区的作风非常严谨，一个patch可能在邮件组中讨论几个月甚至几年，根据大家的意见反复的修正，patch合并到master已经非常成熟，所以PostgreSQL的稳定性也是远近闻名的。

参考

https://commitfest.postgresql.org/13/1054/

https://www.postgresql.org/message-id/flat/CA+q6zcWm_1Ygg5QOq0gYbnB_=zq7G51uexQt3QEgDJa0qQnPKw@mail.gmail.com#CA+q6zcWm_1Ygg5QOq0gYbnB_=zq7G51uexQt3QEgDJa0qQnPKw@mail.gmail.com