Re: Chinese sorting and filtering on an table with charset UTF-8

[Alexander Barkov]

MySQL-5.6.2, which was released a couple of days ago,
includes some collation customization improvements.
You can create a Chinese collation using 5.6.2.

Moreover, the file mysql-test/std_data/Index.xml in the 5.6.2 sources
has an example of Chinese pinyin order:

<collation name="utf8_5624_1" id="354">
<rules>
...
<!-- long tailoring: pinyin order from CLDR's zh.xml -->
<reset><last_non_ignorable/></reset>
<pc>ㄅㄆㄇㄈㄉㄊㄋㄌㄍㄎㄏㄐㄑㄒㄓㄔㄕㄖㄗㄘㄙㄚㄛㄜㄝㄞㄟㄠㄡㄢㄣㄤㄥ>... all 20000+ characters here...座袏做葄蓙飵糳咗</pc>
</rules>
</collation>


So you can just copy its definition and paste into your Index.xml file
with very minor changes:

- Remove the non-Chinese part from the example collation
- Make sure to change collation id to any vacant number
under the range 1-256 (InnoDB currently does not suport 2-byte collation IDs).



Note, earlier MySQL versions will not work, because
the supported tailoring size was limited to only a few kilobytes,
(which was not enough for Chinese).