warning in lookup stage

nair · Post by **nair** » Tue Nov 08, 2005 1:54 pm

Hello Everybody,
I'm getting the following warnings, I'm using lookup stage in the job

Lookup_4: When checking operator: Operator of type "APT_LUTCreateOp": will

partition despite the
preserve-partitioning flag on the data set on input port 0.

Lookup_4: When checking operator: Dropping component "market_vle"

because of a prior component with the same name.

Anybody please suggest me how to eliminate the above warnings

Thanks and Regards
Nair

ray.wurlod · Post by **ray.wurlod** » Tue Nov 08, 2005 2:07 pm

The first message stems from a partitioning requirement of the data - there is not enough information in your post to determine why one might exist - such as your job design. Remove the preserve-partitioning flag; set partitioning to Auto unless you have some reason for doing something different.

The second message stems from the fact that you can not have duplicate column names on an output link. To eliminate this message design the job so that it does not try to force market_vle onto the output from more than one source.

phanee_k · Post by **phanee_k** » Tue Jan 17, 2006 9:54 am

HI ,
I am also getting the same error and not getting cleared if I set the preserve partitioning as clear.

Job design as below.

DataSet
|
|
|
Dataset -------->LookupStage ----------> Transformer -----------> DataSet

I am getting the same error if the source is a sequential file .

How can we resolve this.

Thanks,

ray.wurlod · Post by **ray.wurlod** » Tue Jan 17, 2006 3:13 pm

Exactly the same error?

What are your partitioning settings on both inputs to the Lookup stage?

phanee_k · Post by **phanee_k** » Tue Jan 17, 2006 5:13 pm

Hi Ray,
Partitioning is set to Auto .
It tried with hash partitioning on the reference link and Auto on Primary link.
Tried with setting the preserve partioning to clear.

But no luck.

kumar_s · Post by **kumar_s** » Tue Jan 17, 2006 10:20 pm

Hi,

Try the reference dataset as entire partiton and the main dataset sequence as hash partiton.
If the key in which the dataset is partitoned is different from the lookup key, then you may need to clear the partiton of the dataset and maintain another to lookup.
Also post the warning message you get.

-Kumar

balajisr · Post by **balajisr** » Tue Jan 17, 2006 11:20 pm

Hi

Try this.

Set partioning to Auto for source and reference dataset and preserve partition flag to Auto i.e Let datastage decide about the partitioning and preserve partition flag and verify whether it works.

--Balaji S.R