DSXchange

Tried to explain - it's not a problem with your job. You ran afoul of another process, someone or something else working in the table at the same time you are and locking you out from making your changes. Work with your DBA when it happens again, they should be able to tell you where the lock is com...

The Usage Analysis option is only as good as your developers are disciplined. Metadata must have been stored for all tables involved and that metadata used in the job so that the two have been associated in the repository. Chuck Smith has a utility at his website to find information like you are req...

Why not use the Oracle stage for Oracle then? In other words, why screw with ODBC?

What happens if you make your field a varchar in the job and match the NLS_DATE format of the target database? If you don't know what that is, ask your DBA, they'll be able to tell you how it is set.

Yes, exactly - as I noted.

No, that's not how it works. That option does not mean "wait for the job to finish before running the subroutine" it means "do not run this subroutine if the job does not complete successfully". It always waits for all activity on the canvas to complete before anything 'after job' runs. Something el...

And apparently missed telling us this was only a problem when the job itself failed before you got to the after-job subroutine if that's all it took to 'fix' it.

Dude - people are trying to help you but you are making it difficult by not providing enough information. And bottom line is you need to learn how to help yourself, or at least help us help you. You haven't mentioned the datatype of the Oracle DATE field in the job. You haven't mentioned the version...

Set the Auto Purge that suits each job - X number of runs or X number of days. Do this from the Director. As noted, the Administrator can set the defaults for the Project but that only affects jobs created after the default is in place.

Sounds like you need to create a job to capture this 'lst_updt' and stash it somewhere like USERSTATUS. Then use a Sequence job to run the two Server jobs and pass it as a Job Parameter to the second job. Use the job parameter in the constraint.

Or don't use the 'combo' actions. Split your input stream into discrete insert and update actions.

The 'derivation' in the Pivot stage just becomes a comma delimited list of the column names to pivot, as noted in the documentation.

What does "doesn't always seem to set correctly" mean? It runs without an Invocation ID? Runs with a different one?

And you need to use meaningful subject lines, please.

Search for your actual error message with the bits back in you took out, or perhaps just "Implicit conversion". You'll find conversations on the Parallel Aggregator's appetite for dfloats and what you need to do about it.

For example:

http://www.dsxchange.com/viewtopic.php?t=121324

Pivot pivots, you need to parse. If you know the max # of fields, parse out that single field to the max, pivot and then filter out any nulls. No counting required.