Archives of the TeradataForum

Message Posted: Tue, 13 Mar 2001 @ 09:01:49 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Dave

thanks for the info .

Your assumption that I am trying to improve query time is correct and I understand there will be an overhead on space and maintenace times . The table in question is our main Order table and is the most heavily queried table in the Data Warehouse .

Users would select only on a Date range or only via the week numbers not on a combination of the 2 .

Our marketing / merchandising departments are not interested in financial weeks and would therefore use the date fields and the existing NUSI's give reduced query times .

The finance department need to run selections for financial weeks and thus use the new Week Numbers .

I have estimated that (on average) each week would have approx 600000 data rows and we have a 32 node box thus your calculation gives a value of 2.34. Is this good or bad?.

Thanks to others for their contributions .

Howard .

PS Dave Wellman - could you contact me off list.

	<Prev	Next>		<<First	<Prev	Next>	Last>>