Re: [pgadmin-support] Overloading functions that are used by operators.
От | Dave Page |
---|---|
Тема | Re: [pgadmin-support] Overloading functions that are used by operators. |
Дата | |
Msg-id | 03AF4E498C591348A42FC93DEA9661B83AF172@mail.vale-housing.co.uk обсуждение исходный текст |
Список | pgsql-admin |
> -----Original Message----- > From: Tom Lane [mailto:tgl@sss.pgh.pa.us] > Sent: 10 July 2003 17:07 > To: Dave Page > Cc: Donald Fraser; [ADMIN]; [pgADMIN] > Subject: Re: [ADMIN] [pgadmin-support] Overloading functions > that are used by operators. > > > "Dave Page" <dpage@vale-housing.co.uk> writes: > >> I see no such error here ... > >> > >> regression=# create function pg_catalog.textcat(varchar,text) > >> regression-# returns text as 'textcat' language 'internal'; > >> CREATE FUNCTION > > > Andreas (Pflug) reckons it's gone in 7.4, but in 7.3.2 I see: [ > > complicated query ] > > Oh. I'll bet you're running into the > regproc-reverse-conversion problem. 'textcat' is one of the > histogram values appearing in pg_stats for > pg_operator.oprcode, right? The error comes when the planner > tries to convert that string back to OID form. We fixed this > in 7.4 by modifying the way that pg_statistic stores data > values, so that no conversion needs to occur. Yeah, that sounds like it. The simplified query stops the conversion to oid by just returning the function name rather than joining with pg_proc. Works just as well for us I think and simplifies the query somewhat. Thanks for looking. Regards, Dave.
В списке pgsql-admin по дате отправления: