Michael Meierruth (1/31/2014)
OK, just saw your use of LAG which means SS20012.But you're making completely wrong use of it.
If you delete one of the rows with cgm='5' from your data and run your script the results become a lot worse.
In your final select display the dr column and you'll see this right away.
Also, no dense_rank needed here. Rank is sufficient.
It's not completely wrong - it's getting me all the test data except one row how I want it. Do you have another suggestion?
Switching from dense rank to rank doesn't change output, so I'm not sure what you mean.
Thanks