All Forums General
Vinay151 20 posts Joined 12/14
13 Dec 2015
Insert is taking so much time

Hi, 
  I'm running a insert query which inserts data in 30 iterations. After 16 iterations the record count in the target table is 3 billion. The 17th iteration is taking alomst 7 hours to load the data. Is there any way to speed-up the process time (load time).
 
Thanks,
Vinay.

Bjorn 5 posts Joined 08/07
14 Dec 2015

If you don't have a unique index, try to create the table as a Multiset table. Then Teradata does not need to check for row dupilicates at every insert.

Bjorn S

Bjorn 5 posts Joined 08/07
14 Dec 2015

And check the target table for skewing.

Bjorn S

Vinay151 20 posts Joined 12/14
15 Dec 2015

Thanks for your reply Bjorn. I'm using multiset table only and when I checked for skewing, it's not skewed. 

dnoeth 4628 posts Joined 11/04
15 Dec 2015

Adding partitioning might help.
Can you show the table DDL plus the Insert/Selects?
Why do you need to run so many iterations?

Dieter

Bjorn 5 posts Joined 08/07
15 Dec 2015

Could it be that your last (17.th) iteration has skewed input, with a lot of duplicates of your primary index definition?

Bjorn S

You must sign in to leave a comment.