From: | "David E(dot) Wheeler" <david(at)kineticode(dot)com> |
---|---|
To: | Michael Paesold <mpaesold(at)gmx(dot)at> |
Cc: | Robert Treat <xzilla(at)users(dot)sourceforge(dot)net>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: PATCH: CITEXT 2.0 v4 |
Date: | 2008-07-17 17:07:31 |
Message-ID: | 80B7A7E8-CABD-4663-B13E-2B39324F89CD@kineticode.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Jul 17, 2008, at 03:45, Michael Paesold wrote:
> Wouldn't it be possible to create a variant of regexp_replace, i.e.
> regexp_replace(citext,citext,text), which would again lower-case the
> first two arguments before passing the input to
> regexp_replace(text,text,text)?
Sure, but then you end up with this:
template1=# select regexp_replace( 'Fxx'::citext, 'X'::citext, 'o');
regexp_replace
----------------
foo
(1 row)
Which is just wrong. I'm going to look at the regex C functions today
and see if there's an easy way to just always pass them the 'i' flag,
which would do the trick. That still won't help replace(),
split_part(), or translate(), however.
Best,
David
From | Date | Subject | |
---|---|---|---|
Next Message | Alvaro Herrera | 2008-07-17 17:17:53 | Re: autovacuum crash due to null pointer |
Previous Message | Alvaro Herrera | 2008-07-17 16:42:39 | Re: [PATCH]-hash index improving |