DSXchange

Hi All,

I have a datastage job in which the i am fetching data from a Teradata Connector and second from Dataset.

I join the data using a join stage.

When I am partitioning the data using Hash/Sort method on both the input links, the join is not provioding the correct output.

But when I change the partition method to ENTIRE, the join stage works fine. I cant go for Entire partition as the Volume of data is supposed to be very huge.

Has anyone ever faced such issue before??? If so what can be the resolution as the workaround of ENTIRE does not seem good to me.

Please suggest.

Regards,
Amit

If the inputs are identically hash partitioned and sorted, then it will work. Therefore I must assume that there was some difference between the inputs in your case in how the partition and/or sort keys were specified.

ray.wurlod wrote:If the inputs are identically hash partitioned and sorted, then it will work. Therefore I must assume that there was some difference between the inputs in your case in how the partition and/or sort keys were specified.

We checked it many times, we did not have any discrepancy in the Hash/Sort order of the key columns in both of the links.

Still no Luck. Ultimately we tries the Entire method, where it worked...


What can be the reasonm for this???

As Ray said , if the keys are identical in both inputs, then it will work. Are the data types of the key columns exactly the same as well?

Mike

Hi Amit,

I'm also facing similar issue. How did u use entire partitioning?

My target is a DB2 table. When I give entire partition in join, it gives primary key violation.

Going mad with this... Please help!

If you are using the instage sort then try using a saparate sort stage before partioning, it worked in my case