Re: [QUESTIONS] DANGER WILL ROBINSON!

On Mon, Mar 09, 1998 at 01:02:26PM -0500, The Hermit Hacker wrote:
>
> Redirected to 'the proper list' - pgsql-hackers(at)postgresql(dot)org
>
> On Mon, 9 Mar 1998, Bruce Momjian wrote:
>
> > >
> > > WARNING!
> > >
> > > Postgres 6.3 has MAJOR trouble with btree-indexed text fields.
> > >
> > > Performance levels are *10x* worse than the same indexed fields in "varchar"
> > > format!
> > >
> > > Be EXTREMELY careful - I got bit in the ass by this this morning, and it was
> > > very fortunate that I figured out what was going on.
> > >
> > > The reason I changed this over was that I had dumped the table and it came
> > > out of the pg_dump program with a negative size. So I figured I'd change it
> > > to TEXT and that would resolve the problem. BIG mistake.
> > >
> > > Be on guard for this folks.
> > >
> > > Developers, you might want to look into this - there's no good reason for
> > > this kind of behavior, is there?
> >
> > No good reason at all. As far as I know, text and varchar() behave
> > identically in the backend, except for the input functions which limit
> > the length of varchar.
>
> Karl...just curious, but what does an 'explain' show for the two
> different situations? 10x worse almost sounds like the indices aren't
> even being used, don't they?

Explain claims the indices are being used.

--
--
Karl Denninger (karl(at)MCS(dot)Net)| MCSNet - Serving Chicagoland and Wisconsin
http://www.mcs.net/ | T1's from $600 monthly to FULL DS-3 Service
| NEW! K56Flex support on ALL modems
Voice: [+1 312 803-MCS1 x219]| EXCLUSIVE NEW FEATURE ON ALL PERSONAL ACCOUNTS
Fax: [+1 312 803-4929] | *SPAMBLOCK* Technology now included at no cost