I have table with many inserts, and now DO produce script to insert like this:
but most efficient inserts is something like:
With 25 batch size i checked execution plan: my insert of 25 rows take 8% and each row of first variant takes 3% (3%*25 = 75%) huge difference measured time (with sql profiler) is about 61 ms for single insert and 400ms for 25 separate inserts. |
We're aware of this, but for now we postpone all optimization-related works. Btw, SQL Server supports multirow inserts, this must be even more efficient way. Related issues:
Guys, if this is really important for you, please vote for this question. We'll try to implement this sooner then. |
I think they have some issue with this to support more actions in batches, like DELETE FROM table1 WHERE ID IN (1, 2, 3, ....). Just set lower priority to implement such batches, as i guess.
True: http://code.google.com/p/dataobjectsdotnet/issues/detail?id=538