From: | Robert Vojta <vojta(at)ipex(dot)cz> |
---|---|
To: | Zeugswetter Andreas SB <ZeugswetterA(at)wien(dot)spardat(dot)at> |
Cc: | PostgreSQL Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: indexes and big tables |
Date: | 2001-07-27 13:02:26 |
Message-ID: | 20010727150226.N6382@ipex.cz |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
> The index is only used for the line= part of the where clause
> with your query. With many rows the "line=" is not selective enough
> to justify the index.
I tried move only needed data into new table and change query into ...
netacc=> EXPLAIN SELECT counterfrom AS from, counterto AS to,
floor((980000000 - date_part('epoch', counterstamp)) / 300) AS sequence
FROM graphs_5m WHERE line='absolonll'; NOTICE: QUERY PLAN:
Index Scan using graphs_5m_idx on graphs_5m (cost=0.00..58.38 rows=29
width=24)
EXPLAIN
and query runs for 3-5 seconds. Any idea how to make it faster? I think,
that now it's ready to HW upgrade for faster result ...
Best regards,
Robert
--
_
|-| __ Robert Vojta <vojta-at-ipex.cz> -= Oo.oO =-
|=| [Ll] IPEX, s.r.o.
"^" ====`o
From | Date | Subject | |
---|---|---|---|
Next Message | Leslie | 2001-07-27 14:57:16 | PostgreSQL7.1 on AIX5L is running with too poor ferformance |
Previous Message | Robert Vojta | 2001-07-27 12:57:13 | Re: indexes and big tables |