Re: reset cumulative sum

But you said "I want to reset cumulative percentage on a table if it becomes greater than 1"
IOW you want to recalculate it. That's exactly why it's usually a bad idea to put precentage in the tables. it leads to massive recalculations when the data changes. It suggests to me a design based on a spreadsheet.

And I'm sorry you feel insulted. I criticized your design, not you. Try to step back. your code is not you. Even graduates of the best schools make design mistakes. That's what is called experience.

And no, I didn't insult your school. Since the question you posed was so similar to another recent thread (as Frank noted) I wondered if it was related in some way. people from the same school sometimes tend to follow similar design approaches. Four years with the same CS Professors and clasmates can do that to you.

To close, I will make a suggestion: try a divide and conquer approach. If you can break up the data into distinct sets and have multiple processes working on each set, you may get a performance boost. you will only have to make a slight modification to your current code so the cost to test this is low. If two processes helps, four may be even better. Unless you are in a cluster, 4 is likely the sweetspot in performance, with diminishing gains beyond that.

HTH,
ed Received on Tue Dec 28 2004 - 12:23:38 CST