DSXchange

Hi,

I have Server job with the following flow.

Seq. File Stage --------------->Transformer---------------->Dataset

Please let me know how can i capture the reject records from Seq. File Stage.
In parallel job i can have a reject link coming out of Seq. File stage but since this is a server job its not possible.

Please provide me a logic where i can capture all the reject records which does not meet the metadata of my Seq. File Stage.

Thanks

You can't.

Server edition does not have the concept of data types.

You will need to perform any necessary validation testing in the Transformer stage, and create an output link from that which can handle any rows that are not passed to the third stage.

Server jobs do not have the concept of Data Sets, and therefore don't have a Data Set stage.

Thanks Ray,

My problem is that if say there are 10,000 records in the source file with pipe as delimiter then if there is are few records where pipe is a part of value its causing the whole job to abort.

and it becomes very difficult to identify those few culprit records.

Isn't there a way to separate those records and process the remaining clean records.

bi_fujitsu wrote:Thanks Ray,

My problem is that if say there are 10,000 records in the source file with pipe as delimiter then if there is are few records where pipe is a part of value its causing the whole job to abort.

and it becomes very difficult to identify those few culprit records.

Isn't there a way to separate those records and process the remaining clean records.

As Ray said, you can read the complete line in one column and try to identify those records in transformer and if It doesn't fit your requirement send it to other link as a rejected records.

Regards,

Add a dummy column at the end with "Replace" with blank. If this contains value at any time, reject it in the transformer.

Or you can try this aprroach. If you have 'n' columns, you should be having 'n-1' delimiters(Pipes in your case). Count the number of delimters
using a routine. If there are more than n-1, then reject those records.