parallel inserts for data analysis (ie very large not many small transactions)

  • It is clear that sql 2008 r2 will not automatically parallelize inserts - it will only automatically parallelize the select part of those inserts. Have they improved this in sql 2012 ?

    Large Inserts from one table to another are much faster generally when tablock is used. Is there anything better - maybe something that turns of all logging ?

    It is possible to manually parallelize inserts of course but then tablock wont work so I found I had to use many parallel inserts to equal the speed of one insert with tablock with a typical query. Partitioning did not seem to speed up parallel inserts. Anybody have better ideas on manual parallelization ?

Viewing 0 posts

You must be logged in to reply to this topic. Login to reply