performance... - Mailing list pgsql-general

From chris markiewicz
Subject performance...
Date
Msg-id 000801c08a38$d7b9e5e0$dbb846c6@cmarkiewicz
Whole thread Raw
In response to Re: Calculated values  (Camm Maguire <camm@enhanced.com>)
Responses Re: performance...
Re: performance...
List pgsql-general
hello.

this might be as much of a general database question as it is a postgres
question...

i have a table with 5 columns...a primary key (integer), three small (10
character) text fields, and one semi-large (1400 characters) text field.
note that only a small percentage (5% ?) of the rows contain 1400 characters
in the 5th column...the other 95% have approx 10 characters.  it has 1100
rows.

the problem is this - queries (command line) often take a very long time -
anywhere from 5-15 seconds - to execute.  the queries use only the primary
key and nothing else in the where clause.  no joins.  a sample query is:

select * from weather where weatherid = 12372;

from the command line, it seems that the first query can take a very long
time but subsequent queries happen quickly ( < 1 sec).  i'm guessing that
this is the result of caching or something.

do the long times make sense?  what can i do to shorten them?  would a
smaller text field help?  i have no reason to think that this would be
faster or slower in another db, so it might be unrelated to postgres itself.

i greatly appreciate your help.

chris


pgsql-general by date:

Previous
From: ender
Date:
Subject: converting oracle raw longs
Next
From: adb
Date:
Subject: Re: performance...