DSXchange

Look for dsjob -jobinfo

Will leave the rest for you.

Musn't you be checking why it failed in the first place before executing it again? What if the job was successful last time? You can include this in a sequencer which loops in 10 min intervals (with sleep). You need to get jobinfo and branch on that. Give it a start and let us know if you have any q...

Adding to what ArndW sayd, S9(16)V99 cannot have a value as '1.23'. It will have '123' as V is logical decimal or '12r' (or somethingvery similar) for S.

If the data is in mainframe, are you using DS/390 or something similar? Because server edition cannot connect to mainframe.

Doh!! Me and my typing.

Thanks for pointing Craig

Did you try with some test job?

Reject link will return any rows not loaded into the target - with few exceptions when the datastage with result in core dump

You need to define it.

What have you tried?

Search for define or DSU

Or use a reject link from ODBC stage to capture anything not successful.

Your where clause does not have an '=' symbol in the last line (after col1).

What is the custom SQL statement?

What is the warning / error?

Do not expect others to guess the issues you are facing. It is enough to guess the solution.

or do a sort desc with key change

Ray,

The first stage is an ftp stage (or sequential file). So a transformer connected to this may cause partitioning of data.

I think this is what the op is saying.

But I can't as I do not have funnel in server jobs

But can do link collector or simple "cat file1 file2...."

Btw, what do you mean by 'merge' ?

Did you trying using aggregator and doing max(date) in it?

Maybe you can explain what you do mean by

But i am not able to read all the three files .

You may need an hash file to store the list of values and reference it back.

Alternatively you can do in 2 jobs - one to lookup and write missing records, second to use it together with your original reference file to load the target.