DSXchange

Hi All,

We are having Target oracle tables as hash partitioned.

When using oracle connector can we read tables in parallel ?

When i tried "oracle connector type" partition is used, its throwing out warning cant write and its performing sequential write.

Also , i see "hash partition type" in oracle connector stage, can we use this partition to load parallel ?

Also, instead if i use odbc connector as target does it do parallel write for hash partitioned tables ?

Thank you.

You need SELECT access to a number of Oracle system tables (fully documented in the Connectivity Guide for Oracle), to which DataStage refers to determine precisely in what manner the table is partitioned. You will also need to nominate a partition table property in the Oracle stage.

Oracle hash partitioning is not the same as DataStage hash partitioning, so you will nevertheless need to re-partition (unless keeping the Oracle partitioning is suitable for your purposes).

I understand the accesses needs on system tables.

So , if i do hash partition on a key (same as in table) in connector stage , it will not run in parallel to load table(hash partitioned) since both are different.

The SELECT (if you have specified Partition Table) will run in parallel, the data will be returned partitioned according to Oracle's algorithm.

What you do with the partitioning after that must be guided by your job's stages' requirements.

Ray ,

Thanks for info...Now how to write in parallel if table is hash partitioned . I mean in connector stage what partition i should use to write in parallel(table is hash partitioned) or Enterprise stage.

Thanks.