Problematic Index Scan

From: Tourtounis Sotiris <tourtoun(at)csd(dot)uoc(dot)gr>
To: pgsql-general(at)postgresql(dot)org
Cc: Basilhs Christofides <christop(at)ics(dot)forth(dot)gr>
Subject: Problematic Index Scan
Date: 2002-07-25 15:38:50
Message-ID: Pine.GSO.4.44.0207251822550.10388-100000@amorgos.csd.uch.gr
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

I have a table with 3 attributes(int8, int8, int respectively) and I load
to it a batch file of 150000 tuples, which each of them contains a series
of java long, long and integer parameters respectively too. The first
attribute is classified in descending order and there are indexes on it
and to the third atribute respectively. I make some SQL questions on them
in order to use the created b-trees on the first attribute and for 3-4 times there is
an index scan on it as it used to be. However, after those 3-4 times the
database start to use sequential scan on this first index. Then I stop the
session where i ask those SQL questions, i stop the running postmaster with the
neeeded number of buffers and i except that consecutively
running the postmaster again in a next session there will be again an
index scan on the first b-tree as it was in the beginning. However, index scan
doesn't work in that case as it should be. The only way is to make index
scan work is to drop the whole database where the table is in and re-installed it from the
beggining. I wonder if there is an explanation of that phenomenon and how
i can assure the index scan function without the reinstallation of the
whole table.

SWTHRHS TOYRTOYNHS
(tourtoun(at)csd(dot)uch(dot)gr)

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Tourtounis Sotiris 2002-07-25 15:42:28 B-trees (Indexes) storage space
Previous Message Thiemo Kellner 2002-07-25 15:38:32 Re: CASE Tools