Query on SCD Stage

[balu536]

Hi,
I've a query on Slow Changing Dimensions (SCD) Stage.

Do we need to sort the data (on the input links) on the Business Keys being used in the stage?
Is it mandatory for proper functioning of SCD stage, like we do for Join and Merge stages?


Thanks.

[ray.wurlod]

Data do not need to be sorted on business key. When the dimension table is loaded into memory the business key column is identified. That is sufficient.

[balu536]

Thanks Ray.

Also is it fine if we do partitioning (Hash) on the source link alone or is it required on both source and reference link?

Please clarify.

[asorrell]

Yes, the partitioning should match on both links. However, if the other link is set to "auto" then DataStage will determine that it needs to be set to match the other input stream and set it to "Hash" on the right keys without tell you. I believe if you look at the osh you can verify that.

However, I always recommend setting stages explicitly so it is more evident where the partitioning is occurring.

[asorrell]

Yes and No. Yes, the partitioning should match on both links. However, if the other link is set to "auto" then DataStage will determine that it needs to be set to match the other input stream and set it to "Hash" on the right keys without tell you. I believe if you look at the osh you can verify that.

However, I always recommend setting stages explicitly so it is more evident where the partitioning is occurring.

[balu536]

Thanks Andy.