Re: Review: listagg aggregate
Pavel Stehule <pavel.stehule@gmail.com>
From: Pavel Stehule <pavel.stehule@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Peter Eisentraut <peter_e@gmx.net>, Scott Bailey <artacus@comcast.net>, "pgsql-hackers@postgresql.org Hackers" <pgsql-hackers@postgresql.org>
Date: 2010-01-25T14:12:37Z
Lists: pgsql-hackers
2010/1/25 Tom Lane <tgl@sss.pgh.pa.us>: > Pavel Stehule <pavel.stehule@gmail.com> writes: >> 2010/1/25 Peter Eisentraut <peter_e@gmx.net>: >>> xmlagg -> concatenates values to form xml datum >>> array_agg -> concatenates values to form array datum >>> ??? -> concatenates values to form string datum >>> >>> So it's pretty clear that listagg does not fit into this scheme. > >> when you define list as text domain, then this the name is correct. > > IOW, if you define away the problem then there's no problem? > > I agree that "list" is a terrible choice of name here. "string_agg" > seemed reasonable and in keeping with the standardized "array_agg". > I am not happy, I thing so we do bigger chaos then it is. But it hasn't progress. So I agree with name string_agg. In this case isn't a problem rename this function if somebody would. I'll send patch over hour. regards Pavel Stehule > regards, tom lane >