Re: Select max within max

From: Geoff Muldoon <>
Date: Sun, 13 Jul 2008 10:19:14 +1000
Message-ID: <>

> select colx,
> max(coly) keep(dense_rank last order by coly,colz) ,
> max(colz) keep(dense_rank last order by coly,colz)
> from t1
> group by colx;

That looks groovy.

Many thanks Charles (particularly for doing that performance analysis), Mladen and Maxim.

Rank and dense_rank are the type of techniques I only vaguely understand, but are likely to be the things that I was looking for.

Will sit down and do the analysis on our own data next week with the assistance of my DBA (as I am a mere humble developer) when/if I can get any time in his crowded calendar.

Again, many thanks.

Geoff M Received on Sat Jul 12 2008 - 19:19:14 CDT

Original text of this message