From: | "Merkel Marcel (CR/AEM4)" <Marcel(dot)Merkel(at)de(dot)bosch(dot)com> |
---|---|
To: | <pgsql-performance(at)postgresql(dot)org> |
Subject: | Big array speed issues |
Date: | 2006-06-20 09:35:17 |
Message-ID: | EA04C6394A303A4CADF281DE22F232A41D4357@si-mail48.de.bosch.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
Hi,
I have some speed issues with a big array in a table. I hope you can
help me to tune my query.
My table looks like this:
Id | timestamp | map
Primary key | timestamp | array of real [34][28]
With an index on timestamp
My query is the following:
Select map[1,1], map[1,2] .... Map[34,28] from table where timestamp > x
and timestamp < y order by timestamp
Expected return is about 5000 rows of the table. I have to run this
query multiple times with different x and y values
The table is huge (about 60000 entries) but will get even much more
bigger.
The query takes ages on a 3.GhZ Xeon processor with 2 GB RAM. I'm using
postgresql 7.4 .
Any hints how I can speedup this ? (use postgres 8.1, change table
setup, query one row or column of the array )
I use libpqxx to access the database. This might be another bottleneck,
but I assume my query and table setup is the bigger bottleneck. Would it
make sense to fetch the whole array ? (Select map from table where ...
and parse the array manually)
Thanks for your help.
Marcel
From | Date | Subject | |
---|---|---|---|
Next Message | biuro | 2006-06-20 12:39:32 | Curson prbolem |
Previous Message | Luke Lonergan | 2006-06-20 01:24:02 | Re: Some performance numbers, with thoughts |