Partition Table option in ORACLE enterprise stage

[ragavendradineshdsx]

Hi all,

I have a question regarding the usage of "Partition Table" option in ORACLE enterprise stage. I have a job with 3 stages
(ORACLE stage --> Xformer --> Dataset).
I'm reading 4 rows from the table A and just dumping it in dataset.I'm using a 4-node configuration file. When i run this job by not using "Partition Table" option in the oracle stage (i mean, in sequential mode), then i have 4 rows distributed in all 4 nodes equally
(i.e, node1->row1,node2->row2,node3->row3,node4->row4).
But, when i use the option "Partition table" in the oracle stage and run the job, i'm able to see all 4rows being dumped into a single node (in my case, all 4 rows are dumped into node2). I'm really not sure why this is happening? FYI, i'm using "AUTO" partition method in all stages.Is this behavior expected? or, am i doing something wrong in the job execution?
Let me know your thoughts on this!

[jwiles]

The Partition Table option allows the operator to read from Oracle in parallel based on how the table you specify is (or ISN'T) partitioned in the database. The operator is running in parallel execution mode when you specify this option, unless you've forced it to run in sequential mode. As the next stage in your job is likely running in parallel, the data leaving the oracle stage will not be repartitioned by 'Auto' unless Orchestrate sees a need to do so.

When you don't specify the Partition Table option, the operator runs in sequential execution mode. As the next stage in your jobs is likely running in parallel, the data leaving the oracle stage is probably being round-robin partitioned.

Do you know if the table you are specifying is partitioned within the Oracle database? If so, all four records may be in the same table partition. If not, they definitely are (there's only one partition).

Regards,