DSXchange

Posted: **Thu Dec 16, 2010 10:37 am**

(split from this topic)

DSguru2B wrote:Make sure your hashed file is created and read with identical metadata. Also make sure your your keys are trimmed and both source and reference dates are in identical formats.
Another thing to keep in mind is that your input data may have duplicates as well, you need to handle that along with the duplicates in your target table.

Hi,

Could anyone please let me know if its possible to perform a sparse lookup in server edition..if so how?

many thanks,

Posted: **Thu Dec 16, 2010 1:37 pm**

Sparse lookup is a direct lookup to the table without any use of hashed files. Used only if your source record is small. Sparse lookup come with a performance hit.

Posted: **Tue Dec 21, 2010 2:19 pm**

I'm still confused here..

Is it like this?

if my source is a primary link to the transformer and target is db..then use replica of the target as reference link to the transformer?

Posted: **Wed Dec 22, 2010 4:34 am**

Posted: **Wed Dec 22, 2010 12:13 pm**

If I use hashfile stage between the Db and the transformer for the reference link does that become a normal lookup? because in parallel db stage there is option available to chose between normal and sparse. I don't really know if that concept applies in server jobs as well.

Posted: **Thu Dec 23, 2010 1:07 am**

Hi,

Yes. It becomes normal lookup not the sparse. Sparse Lookup means using reference link between DB stage (it may Oracle OCI or any other stage which is access database directly) and transformer. In this case Hashfile stage is not used.

Posted: **Thu Dec 23, 2010 11:39 am**

Thanks everyone!

DSXchange

Sparse Lookup in Server

Sparse Lookup in Server