The fastest way to get a single record May 22, 2009
Posted by mwidlake in performance.Tags: performance
trackback
Piet de Visser and I have had the odd chat about SQL performance over the last couple of months and he recently posted this interesting article and test script showing how to get a single record out of a large table with only one logical IO. It uses hash clusters, which are a very underutilised feature.
I’ve a couple of thoughts myself on how to use them, but for now I’ll leave you to look at Piet’s posting.
Martin Widlake's Yet Another Oracle Blog 
Thx Martin. In between a major culling, an upcoming additional foreign project, and some holiday days, I am trying to follow up on clusters. But the more I read and experiment on a topic, the less I know about it. And coneveying findings in clear, usable form to other is an art as well.
Feeling humbled by oracle at the moment, but will persist.
But Yes, I do think IOTs and (hash)Clusters are under-utilized. But then an IOT bug nearly took my career away from me (bug 6646613, and ML 471479.1). Ouch.