how to find out what a partition a dataset uses

ruf888 · Post by **ruf888** » Thu Mar 08, 2012 7:17 am

Hi, all

I'd like to know if there is a way to find out which partition a datasets is using ?

BI-RMA · Post by **BI-RMA** » Thu Mar 08, 2012 7:52 am

Persistent in Job log:
Include Variable APT_RECORD_COUNT and set to true. This will report how many rows have traveled which links per node.

In Director job monitor:
In Director when You open a Job Monitor You should see the number of nodes the job used during the last run. Right click on a stage that shows something like "<Stage name> x2" and activate "Show Instances". After that You can see the link statistics per node.

ruf888 · Post by **ruf888** » Thu Mar 08, 2012 10:17 am

Hi, Roland,

I think you misunderstood me. I'd like to know if the datasets is partitioned by which key, not how many nodes it uses.

For example, if I partitioned the datasets by type/model, where can I identify that this ds is partitioned by type/model except in the job itself.
If the job always uses the same partition, then must go back several jobs to find out which partition key it has used.
Is there any way to find it out without lookingat the jobs ?

Jboyd · Post by **Jboyd** » Thu Mar 08, 2012 11:44 am

I believe setting the APT_DUMP_SCORE environment variable to true will assist with this.