[groonga-talk:323] [ANN] PGroonga 1.0.9

[Kouhei S. <ko...@cl...>]

Hi,

PGroonga 1.0.9 has been released!

  http://groonga.org/en/blog/2016/06/02/pgroonga-1.0.9.html

PGroonga is a PostgreSQL extension that makes PostgreSQL
fast full text search platform for all languages!

Here are changes since 1.0.6:

  * Supported PostgreSQL 9.6beta1.

  * [Ubuntu] Supported Xenial Xerus (16.04 LTS).

  * Added pgroonga.highlight_html function that returns
    search keyword highlighted HTML.

  * Added pgroonga.match_positions_byte function that
    returns locations of keywords in text.

  * Added pgroonga.query_extract_keywords function that
    extract keywords from query.

  * [Windows] Upgraded bundled Groonga to 6.0.3.

  * [Windows] Upgraded build target PostgreSQL to 9.5.3.

  * [pgroonga.text_array_term_search_ops_v2 operator class]
    Added &^> operator that performs prefix search against
    text[] type value. If any element is matched, the value
    is matched.

  * [pgroonga.text_array_term_search_ops_v2 operator class]
    Added &^~> operator that performs prefix RK search
    against text[] type value. If any element is matched,
    the value is matched.



This release adds pgroonga.text_array_term_search_ops_v2
operator class. You can use prefix search and prefix RK
search against text[] type with this operator class. They
are useful to implement input completion on search text box.

The following description shows how to use prefix search and
prefix RK search. The following description uses an example
that implements tag name input completion.

First, you need to insert tag names and tag readings. Tag
readings should be in Katakana.

----
CREATE TABLE tags (
  name text PRIMARY KEY,
  readings text[]
);

INSERT INTO tags VALUES ('PostgreSQL', ARRAY['ポストグレスキューエル', 'ポスグレ']);
INSERT INTO tags VALUES ('Groonga',    ARRAY['グルンガ']);
INSERT INTO tags VALUES ('PGroonga',   ARRAY['ピージールンガ']);
INSERT INTO tags VALUES ('pglogical',  ARRAY['ピージーロジカル']);
----

You need to create indexes against tag names and tag
readings. It's important that
pgroonga.text_array_term_search_ops_v2 operator class is
used for tags.readings.

----
CREATE INDEX pgroonga_tags_index ON tags
  USING pgroonga (name pgroonga.text_term_search_ops_v2,
                  readings pgroonga.text_array_term_search_ops_v2);
----

Here is a SELECT to use prefix search against tag names such
as PostgreSQL and Groonga:

----
SELECT name
  FROM tags
  WHERE name &^ 'pos';
--     name    
-- ------------
--  PostgreSQL
-- (1 row)
----

Here is a SELECT to search tags by romanization of Japanese:

----
SELECT name, readings
  FROM tags
  WHERE readings &^~> 'pos';
--     name    |             readings              
-- ------------+-----------------------------------
--  PostgreSQL | {ポストグレスキューエル,ポスグレ}
-- (1 row)
----

You can use OR to get both results:

----
SELECT name, readings
  FROM tags
  WHERE name &^ 'pos' OR
        readings &^~> 'pos';
--     name    |             readings              
-- ------------+-----------------------------------
--  PostgreSQL | {ポストグレスキューエル,ポスグレ}
-- (1 row)
----

Here is an example that searches by pi-ji-:

----
SELECT name, readings
  FROM tags
  WHERE name &^ 'pi-ji-' OR
        readings &^~> 'pi-ji-';
--    name    |      readings      
-- -----------+--------------------
--  PGroonga  | {ピージールンガ}
--  pglogical | {ピージーロジカル}
-- (2 rows)
----


Thanks,
--
kou