Sunday, July 3, 2011

KScope 11: Top Tips to Get Optimal SQL Execution All the Time

This was my final session of the conference.

Prior to this one was Jean-Pierre Djicks on Managing Parallel Execution without Tuning in 11g Release 2. Conveniently, it was in the same room. Especially convenient given that it was in the same room the day after the Queen Mary event. I did want to see JP's other talk, Speed Up Your Data Warehouse with in-memory Processing, but it was just too much effort for me.

Besides, I had never seen Maria speak and I was trying to broaden my horizons. I wish I had seen her speak years ago...would have made life a lot easier.

Maria is not on Twitter but you can find her writing at the Oracle Optimizer blog.

The optimizer subject has always been a bit intimidating to me. I don't know why, it just has. I've mostly glossed over it.

No longer.

Despite my state (post Queen Mary event), a lot of things came together. The optimizer really isn't magic. Surprise!

The importance of statistics, which I've always known but never truly understood, was made clear. Extended statistics? Neat. Histograms, explained (better).

Awesome stuff...and really not as magical as you might think. Take a gander at the Optimizer group's Explain the Explain Plan for more details.