Gustav Brock
Gustav at cactus.dk
Thu May 1 11:50:44 CDT 2008
Hi Arthur If you have timestamps on your records, select current data between last timestamp of the aggregated data and chosen timestamp (now) and union these with the aggregated data. /gustav >>> fuller.artful at gmail.com 01-05-2008 18:10 >>> This is becoming an important issue for me. Assume an OLAP db that gathers its data with a frequency of let's say once a day, and an OLTP db that gathers its data very frequently, let's say 100 transactions an hour. I need to combine them into a result set. While we're all asleep, the OLAP part can roll up its aggregates etc., that's not complex. But I'm wondering about the joins with the current data plus the history. I am not sure how this is done. Never done it before, want to get it right. The OLAP part is simple: dimension this and that and the other thing and you're done. But the current data happens way too quickly to think about rolling it into the OLAP part and then interrogating the cube. So how does one combine the cube with the OLTP data? I've been reading this and that and so far have no clear-cut answer on this. Anybody been there and done that? TIA, Arthur