Monday, January 04, 2016

A Lot of Extraneous Data Sets?

In a recent blog post here I talked about a quick and dirty method of converting your partitioned table spaces from index-controlled to table-controlled. If you haven't read that post, take a moment to click over and read it here: Easily Convert to Table-Controlled Partitioning.

The reason I bring this up today is that I received an interesting e-mail from a long-time friend and DB2 DBA who read the post and had some information to share with me. He told me about how his organization used one of my tips to drop unused indexes as part of this process.

He told me that during the conversion process they dropped a lot of the clustering indexes because they weren't being used for access paths or for uniqueness. And they were able to release an "astonishing 4,100 data sets" by doing so!

Now I am not suggesting that every shop will be able to experience a similar savings, but if you have indexes that have no purpose other than enforcing index-controlled partitioning, it is time to bite the bullet and drop those indexes as you convert to table-controlled-partitioning (and then on to Universal table spaces).

And when you convert, please drop a note here on the blog to let us know how your conversion efforts went!

No comments: