From: | Alexey Klyukin <alexk(at)commandprompt(dot)com> |
---|---|
To: | Haifeng Liu <liuhaifeng(at)live(dot)com> |
Cc: | "pgsql-admin(at)postgresql(dot)org" <pgsql-admin(at)postgresql(dot)org> |
Subject: | Re: Does index based on text with 'en_US.utf8' collation support like query? |
Date: | 2012-12-25 12:44:58 |
Message-ID: | F86B9A85-59E6-4577-856A-E6EE10B0D352@commandprompt.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Hello,
On Dec 25, 2012, at 1:37 PM, Haifeng Liu <liuhaifeng(at)live(dot)com> wrote:
>
> Hi,
>
> I found that queries like "select * from tbl where col1 like 'abc%'" will use index only when the col1(and the index) use 'C' collation. If I use 'default' which is 'en_US.utf8' in my database, the equation condition will use the index but like query will not.
This is a documented behavior, LIKE with a default operator class only works with C locale.
>
> I also tried to specify the collation explicitly: like ('abc%' collate 'en_US.utf8'), not work too.
>
> Is that only 'C' collation support 'like' query or I missed something in my sql? Or do I need to install something additional to support index on other collation?
You need to add a new index with a special operator class called 'text_pattern_ops' (or 'varchar_pattern_ops), i.e:
CREATE INDEX col1_pattern_idx ON tbl(col1 text_pattern_ops);
This index will be used in LIKE queries under non-C locale. You also need to preserve your existing default (aka text_ops) indexes if you want comparison or inequality operations on the col1 column to be indexed.
Refer to http://www.postgresql.org/docs/current/static/indexes-opclass.html for more details.
--
Alexey Klyukin http://www.commandprompt.com
The PostgreSQL Company – Command Prompt, Inc.
From | Date | Subject | |
---|---|---|---|
Next Message | Devrim GÜNDÜZ | 2012-12-25 16:00:31 | Re: Setting up High Availability Group of PostgreSQL instances using RedHat Cluster Suite |
Previous Message | Haifeng Liu | 2012-12-25 11:37:10 | Does index based on text with 'en_US.utf8' collation support like query? |