July 15, 2014 at 12:10 am
Comments posted to this topic are about the item Implementation of partition switching within SSIS
[font="Tahoma"]Cheers,
Sean :-D[/font]
July 15, 2014 at 2:23 am
Regarding running multiple load processes into a partitioned table at the same time..
I have found that most systems have a optimal degree of parallelism which is influenced by number of CPU cores (as you discovered), amount of RAM and disk/storage contention.
It's worthwhile running the system with varying degrees of parallelism to find which one works best.
July 15, 2014 at 6:48 am
Very nice article. Thanks!
July 15, 2014 at 7:33 am
Why drop and create tables, shouldn't 'truncate' command do the same thing with less overhead?
July 15, 2014 at 10:56 am
The previous articles in the series describe why we can't use the truncate and reload scenario in detail. The short story is that this is a multi-tenant system and we don't want to clear out the other customer's data (From previous loads in the past and earlier in the same day). Also for our large fact tables the data is too large to load for all customer's completely daily (which is what a truncate and reload would require). For some of the fact tables the source system doesn't carry history, so we could not do a full reload every day even if we wanted to.
[font="Tahoma"]Cheers,
Sean :-D[/font]
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply