Sequential File Rejects

rajeshknl · Post by **rajeshknl** » Tue Sep 16, 2008 5:23 pm

Hi

The sequence of the stages in my job is as below

Seq_File_Source--->Transfomer--->Lookup--->Seq_File_Target

Not even a single record is getting inserted into my target sequential file.

Selective text from job log of director which might be helpful

1)FIN_HDR,0: Import complete; 516 records imported successfully, 0 rejected.
2)TR_LKP,0: Output 0 produced 516 records.
3)LKP_HDR,0: Input 0 consumed 435 records.
4)LKP_HDR,0: Output 0 produced 1 records.
Output 1 produced 0 records.
5)LKP_HDR,0: Output 0 produced 516 records.
6)HDR_TARGET,0: Input 0 consumed 516 records.
7)HDR_TARGET,0: Export complete; 0 records exported successfully, 516 rejected.

I dont get why all the records are rejected.
I tried having a reject link and write to a reject file.
But even then they are getting rejected at the reject file instance.
That makes sense because both cases are file targets and have the same metadata.

I do the exactly same lookup in several other jobs and they all run fine.

Is there any other location(other than director) where i can find detailed reasons why datastage/file is rejecting these records?

I used to use Informatica earlier the log gives in very detail about any rejects. Datastage just says rejected does not give any reason.

ray.wurlod · Post by **ray.wurlod** » Tue Sep 16, 2008 8:13 pm

There's only one reason for rejecting when exporting, and that is that the data do not match the metadata. Add a Reject output link from HDR_TARGET to capture the rejected rows, write them to a text file (as a single string) to see what is happening.

Actually, all rows can also be rejected if you don't have write permission to the file, but I assume you've already checked that.

ambasta · Post by **ambasta** » Sat Sep 20, 2008 3:04 am

If you have null in the target record..It will be rejected from the file

ambasta · Post by **ambasta** » Sat Sep 20, 2008 3:04 am

If you have null in the target record..It will be rejected from the file

ray.wurlod · Post by **ray.wurlod** » Sat Sep 20, 2008 4:14 am

That is simply incorrect.

However it is necessary in the Sequential File to state how NULL is to be represented in the file. This can be done globally (for all fields) or on a per-field basis.

rajeshknl · Post by **rajeshknl** » Sat Sep 20, 2008 12:35 pm

I intially developed the job with a Netezza Enterprise target. I had to disable RCP as Netezza Enterprise was failinig because of it.

Then i changed the target to sequential file and all the records were beig rejected. I re-enabled the RCP and the job is running fine.

IBM guys say that enabling/disabling RCP causes the compiling to be different and that might be the issue.

ray.wurlod · Post by **ray.wurlod** » Sat Sep 20, 2008 3:44 pm

It is not strictly the compiling that is different, it is the code that is generated that is different.