DSXchange

Try double-quoting the filename with spaces, as the operating system will read space seperated strings as more than one argument. As for the filename - what are you wanting to do with this exactly? You can add an option the stage to include 'file name column' which will add the filename the data was...

Cheers Arnd, with you now... I forgot it was a 'list' approach. Brain hasn't restarted following bank holiday weekend yet

Ta

Sainath, isn't that the requirement - to replace ". " with "."

As in Convert(". ",".",string)

Is there something wrong with using Convert for this then?

In stage Transformer_145 are you calling a routine named "DateVal" par chance? It seems this routine has issues of its's own...

How many records are being processed through this job (including on each side of each join)?

Add environment variable $APT_DUMP_SCORE with value True and rerun the job, then copy and paste the output from the job log here please.

So what was the issue caused by the policy number? Was it a null / a zero?

Convert string function - the product documentation has a functions reference that lists all those available by datatype.

OK, that suggests the julian date conversions have a specific issue then. I'd raise it as a PMR through your appropriate support lines.

Convert/read the second date field to/as a string, use a date from string conversion to achieve consistency across the two and then perform your comparison.

You say you have previously combined explicit conversions in a single specification - correct? Which conversions did these use? I've tried one as below which fails with a similar error as yours: date_out:date=date_from_julian_day(julian_day_from_date(date)) ... which suggests its either not possible...

Check all is well on the database side i.e. manually connect and run the query. Also, as before, are there any other indications of a problem in the log? Does the job monitor show a particular stage in unknown or aborted state?

Are there any other warning or fatals in the job log other than that entry? Even go as far as checking every other entry in there, regardless of type.

What's the job design?

Any reason why you are using an ODBC stage rather than a native DB2 stage?

EDIT - I just spotted in your post that the warning illustrates you are connecting to mainframe DB2.... correct?