| From: | "Thomas T(dot) Thai" <tom(at)minnesota(dot)com> |
|---|---|
| To: | <oleg(at)sai(dot)msu(dot)su> |
| Cc: | <alvherre(at)dcc(dot)uchile(dot)cl>, <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: string || NULL ambiguity |
| Date: | 2003-03-20 02:56:24 |
| Message-ID: | 2060.63.226.186.156.1048128984.squirrel@mail.minnesota.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> On Wed, 19 Mar 2003, Alvaro Herrera wrote:
>
>> On Wed, Mar 19, 2003 at 10:26:05PM +0300, Oleg Bartunov wrote:
>>
>> > we have a little problem in new version of tsearch we're currently
>> working. We've implemented concatenation operation for txtidx type
>> and treat concatenation with NULL as NULL (as it should be).
>> > But people get confused with such behaivour. Do we obliged to
>> > follow NULL rule ? It seems more natural in case of text searching
>> to treat stringA||NULL as stringA.
>>
>> Why don't you just enclose the possibly NULL fields in COALESCE?
>>
>> stringA||COALESCE(NULL, '')
>>
>
> we don't know in advance if it's NULL or not.
I'm the user who reported this on the OpenFTS list. May I suggest allowing
'' in txtidx column type as we do with datatype text, char, and varchar.
That way concat will work as expected for text searching.
--
Thomnas
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Bruce Momjian | 2003-03-20 02:59:03 | Re: cursors outside transactions |
| Previous Message | Bruce Momjian | 2003-03-20 02:42:05 | Re: Open 7.4 features |