On Thu, 8 May 2003 09:20:31 +0200 (CEST), <g.hintermayer@inode.at>
wrote:
>So that's a factor of about 10 faster, only by changing the user, very
>strange.
Very, very strange! Compare the outputs of SHOW ALL for both cases.
If there are any differences, please inform us.
> -> Seq Scan on produkt (cost=0.00..2417.41 rows=2141 width=40)
^^^^ ^^^^
> (actual time=0.02..27.12 rows=2141 loops=1)
Unless I'm missing something, your produkt table has more pages than
tuples. VACUUM FULL should reduce its size to ca. 22 pages.
Servus
Manfred