DSXchange

Set "enable partioned read" to true in the ODBC Connector. Then choose between a modulus or range partitioning method.

Are you sure that you actually have a microsecond value in the timestamp from the other two source?

Try to post examples of all three source timestamps.

Do you want the timestamp truncated as in all zeros (.000000) or as in not shown at all?

What length and scale have you set for the timestamps?

You can join the two output from the sequential file stage on the key value you are generating.

hmmm... we use HP Quality Center too....

I know! Just didn't know what else to call it...

You can easily use the sequential file stage too...

Set the NLS (IBM01142 or check with the mainframe admin)
Set the format (fixed length, binary, Packed decimal)

It should not make a performance difference either way If you are looking to get a performance improvent for your job (besides running on more nodes), then you could check whether you are able to read parallel from your source. If you want to know whether is it good developing practice to set partit...

But using the GUI for changing names of jobs that are being invoked in a sequencer is rather a big pain if you map a lot of parameters. Reason: You can't change the name of the job in the job activity stage, but will need to browse for it, and then all parameter mappings disapper even if you new job...

Your job logs can end up taking a lot of space if no purging options have been set, and this can in my experience result in some the errors you mention.

Are you sure you have set the entire partitioning method on the right input link? (there is a drop down of all input links)

We have experienced the problem a couple of times (two different developers), but with no luck in identifying a root cause.

Solution: re-installation of the client

yes - in that case you'll still be sending records with null values out the link, which causes the runtime error you are describing.

So, map the stage variable to your derivation for cocpysno...

You write you are performing your null handling in a stage variable - so do you make sure to either use the result in a constraint on the output link or as input to your derivation?

As default only before routines fail/abort on error code = 1, where as after routines only write a warning to the log.

So - maybe you can write your own version of ExecSH (I'm guessing that it is the one you are using?) where you abort the job on failure (still just a best guess).

We map the CLOB datatype to NVARCHAR(MAX) (on SQL server 2008 R2)