DSXchange

They will be skipped when locked / open however. And IMHO you should stick with dsx for whole project exports. I find xml clumsy and the first thing DataStage is going to do with an xml export when you go to import it is convert it to dsx first, so you can save a (potentially quite long) step by usi...

satish.raghavan wrote:Can anyone help me because hanging around with this problem for a long time.

Why not involve your official support provider rather than 'hanging around' with it?

Thanks for the explanations.

Any 'core' files would need to be handled by the O/S, probably a simple cron script to find and nuke them. And 'log files' means job logs? If so, you handle them with appropriate Auto Purge settings.

Me, I know I'm not Wright. :wink: If you don't have duplicates in the incoming stream, or you don't need to record what keys you just inserted, you can skip the whole 'update the hashed file' part. Or for the latter you can defer the update by dumping new keys to a flat file and loading them back in...

Getting closer. :wink: Between postings I tried the sort stage not sorting and I got the following fatal error when I ran the job: Sort_16,0: Input not sorted at record 54332 This is untrue unless it was somehow an artifact of the partioning (running on 4 nodes). There is a Sequential File stage rea...

OK. I had already tried a Sort stage (set to Don't Sort, Already sorted) before the Aggregator and it didn't seem to do the trick. However, I hadn't changed the Partitioning / Collecting option inside the Aggregator Input tab from the default of Auto. Do I need to Hash on the grouping key and then c...

OK, what if my data is already sorted? Is there any equivalent to the Sort stage's "Don't sort, already sorted" functionality?

Another PX Aggregator question - when grouping and summing, it is possible to get it to output records on a group change rather than hold onto everything and only output once all inputs have been processed? Looking for the equivalent on the Server side where you can assert that the input is sorted b...

DRS set to Oracle should generate the same contraint violations, suspect you mean via ODBC, yes?

You need a lookup for that so that you don't even attempt the insert if the row already exists. That or use a combo action so when the insert fails an update can be done.

An exit status of zero would indicate success.

Can you post the complete log entry that shows the Execute Command stage running and what it returned?

gssr wrote:Do i have to mention Char(9) for it?

Yes. And that is the resolution, not a 'workaround'.

People will be happy to 'guide' if you have specific questions but no-one is going to type up a complete lesson, for that there's the documentation and Help button inside the stage. However, you will not be able to use the FTP stage with anything that's not a plain old sequential file - i.e. no RAR ...