Issue in record Extraction

[ag_ram]

Hi Folks

Facing a strange problem , extracting data from z/os - DB2 . Link counts show 10000 records and view data shows 10000 , actual dump into a dataset has only 100 records immaterial of data extracted !

[ray.wurlod]

How are you determining the number of rows in the Data Set?

What else is happening in the job design? Is there anything at all in the job design that might prevent records from being passed through? Is the job compiled in trace mode?

[ag_ram]

Hello Ray

I am using DataSet-Management utility to read the dataset and the no:of records in each node.
There are no other stages in the Job , it is just a dump of the table we are taking for processing in the subsequent Jobs.
Job is compiled normally not in trace mode

[balajisr]

Is there any warning in the log?

[ag_ram]

NO Warnings / Errors in the log

[ArndW]

Are you sure that you are looking at the actual record counts and not some limit placed by the "view data" in the Data Set Management tool?

[sas]

Create a small dummy job to read the same dataset instead of reading from Data Set Management tool.

[sas]

Create a small dummy job to read the same dataset instead of reading from Data Set Management tool to check how many records are being read.

[ArndW]

I like using the command line tool orchadmin ll {dataset} ; that gives me the number of records per partition, total bytes, etc.

You could try

Code: Select all

orchadmin dump {DataSet} | wc -l

to get a 2nd opinion on the number of the records in the dataset.

[ag_ram]

HI Folks

Tried both ! Still remains unresolved though .

[balajisr]

Try writing to sequential file instead of dataset.

[ag_ram]

No still does not work Still

[JeroenDmt]

Did you check the data in DB2? Is 10000 the correct number, or are the 100 records that you are getting, the correct result?