DSXchange

Timestamps has colon (:) symbol which is generally interpretted for remote hostname or ip address.

Parameter must not be computed within the job.

Use an afterjob rountine for now.

Maybe the keys in the stage are more than you need.

Maybe you can filter them out in a stage or do 'grep -v' for something special in header.

As Craig noted, the job information is distributed across multiple objects.

Also the size of log file will vary depending on number of runs, warnings etc.

Maybe you can count the stages in the job to determine its size - logically speaking !!

Even though you can, it will leave a mess around - especially if you try to read any.

So best is to use "orchadmin rm".

A simple wil do the trick.

or use %dd%mm%yyyy or %mm%dd%yyyy format (US or UK) to your source date rather than splitting.

Sourt -> Sort -> Transformer -> Output

Sort by key1
Check all columns with previous using stage variables in xfm.

You can set commit size to 0 which will make it everything or nothing.

Select the record where flag='YES' and do a Getlinkinfo in the output link.

Are you saying that the columns are not ordered as you expect.

This depends on the way you created it. For example, do you use RCP?

What is your plan for ?

Did you delete a project with same name before ? Or perhaps aborted the creation ?

Even though trigger can be another design, I doubt it will increase the speed efficiency as it will be another SQL fired for each row. Alternative solution may be to identify currval of sequence number before and after the run and use it to obtain list of inserted values. This is just a starting poi...

If the input brings unique keys, you can reverse the main and reference streams.

Did you compile ?