use of peak stage

chrisjones · Post by **chrisjones** » Fri Jun 21, 2013 8:16 am

Hi ,
Can someone help me for below requirement .

I am getting data in below format from source

col1,col2,col3, col4

1, 100 , , xyz
1, 100, abc,

the expected output should be

col1,col2,col3, col4
1, 100 , abc , xyz

i know we can accomplish using the peak stage but not sure how to do it.

prasson_ibm · Post by **prasson_ibm** » Fri Jun 21, 2013 8:38 am

Hi,
Its not "Peak"stage ,but it is "Peek" stage.

Peek stage is only used for debugging,you can see the output of peek stage in datastage director log.

miwinter · Post by **miwinter** » Fri Jun 21, 2013 8:42 am

Prasoon,

You can also use Peek to pull out specific columns/values to send to an output link, from which you can join it back to something else, for example, or do some other processing on. It's not restricted solely to being output to the job log.

ray.wurlod · Post by **ray.wurlod** » Fri Jun 21, 2013 3:41 pm

I would use Transformer stage for this requirement, either using LastRecordInGroup(col1) as a constraint expression or following the Transformer with a Remove Duplicates stage set to "Last" and grouping on col1.
Within the Transformer populate the column if the incoming value is non-empty.