db-derby-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Knut Anders Hatlen <knut.hat...@oracle.com>
Subject Re: Indexing speed in presence of BLOB columns
Date Fri, 07 Jan 2011 08:28:27 GMT
Trejkaz <trejkaz@trypticon.org> writes:

> Surprisingly, the larger the data which is present in the BLOB column,
> the longer it takes to index.  Since the indexing process shouldn't
> need to read the BLOB data in order to index the row, I am surprised
> that it slows down when more data is present.  What is going on here
> exactly?

Hi Daniel,

You're probably seeing the effects of this issue:
https://issues.apache.org/jira/browse/DERBY-1506

> Is there some way to avoid this (which won't require changing the
> schema)?

Not that I'm aware of. Except if you create the index before populating
the table, you'd avoid the table scan and the unnecessary reading of the
BLOB columns. But that's not an option for existing tables.

-- 
Knut Anders

Mime
View raw message