db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From nigelm <nigel.mag...@gmail.com>
Subject Large IN clause performance
Date Sun, 03 Sep 2006 14:00:42 GMT

Hello

I have a dataset that's too big to be viable for HSQLDB, so I loaded it into
derby instead.

The existing code makes a lot of use of SQL "IN" clauses, where the term can
often contains thousands of items.

This performs very badly on 10.1.3.1 - because it seems to ignore indexes
and just use the min and max terms to do a table scan - this isn't going to
be good as there's millions of rows in there. In fact, even if the IN clause
contains only a handful of entries it goes quite slowly.

Is this something that's addressed in a later version (should I bump to a
pre-release), or am I stuck with either re-writing to get better
performance?


-- 
View this message in context: http://www.nabble.com/Large-IN-clause-performance-tf2210872.html#a6122846
Sent from the Apache Derby Users forum at Nabble.com.


Mime
View raw message