DSXchange

That's the first mention you've made of that fact!


What error messages are logged by the aborting job?

What is the data type of the YEAR field? My guess is Char(4).

Now, what is the Null Field Value property set to (a) for the YEAR field or (b) generically?

Try the grep at the command line. It does NOT return the count. You would need to pipe the output of grep into a wc -l command to get a count.

You will be able to do that once Dean advises whereabouts in the RT_CONFIGnnn record the schedule information resides.

Make the output fields nullable.

1.) Yes - use the Schema File property. But you need to provide a schema file at run time that matches the structure of the data file.

2.) Yes. Read it using a single VarChar field and parse that within the job.

You can do it without conversion if the date strings are in a sortable format, such as YYYY-MM-DD.

Something like

You can not do this in the Lookup stage itself (one stage, one task) but can do so in any downstream Modify or Transformer stage. Pass the columns from the reference link to the output and replace null with -99 as required.

Your field width is defined as 4 and therefore whatever you have specified as the null field value must have precisely four characters. Apparently what you have specified for null field value contains five characters.

True if you have only one File property. Try it with two File properties.

Have a reject-handling output on the Transformer stage. If used, this will log a non-informational message in the job log, which can be detected by the job sequence since the job's exit status will not be DSJS.RUNOK.

What does the grep command generate as output? What are the record schemas going in to and out of the External Filter stages and the actual grep command used?

No, I mean manually when designing the job.

Would your glasses be half full or half empty?