DSXchange

Hmmm... the link I posted worked (and still works) fine for me. Looks like you found the same information, however.

Hopefully this helps.

You want to count the number of columns that do not contain nulls or do something with the counts from inside all non-null columns? Seems to me you'll still have to check each one individually for null so I'm thinking just stage variables in a transformer will be appropriate.

It cannot be a simple NUMBER. Is this actually a TABLE OF / INDEXED BY data type? Regardless, unless it's just numbers stored in a delimited string field you'll have to do the extraction of the components in the database. And I have to ask - any particular reason you put a "{" on the left ...

Andy, I think so but it seems to me in that case it would never work properly rather than not work once in a while. And Tony, you can actually look at the generated Job Control code in the Sequence job to see what / how it checks the status. Actually, you may be thinking of a Reject constraint, I do...

Have a chat with your DBA. It needs to maintain a 'snapshot' of the data's initial values over the period of time your select happens and it uses UNDO space for that. Other processing on the system is eating it up and thus it cannot maintain your snapshot. So as you noted, one solution is to add mor...

TonyInFrance wrote:when we define a trigger as Expression Type is OK - (Conditional) does this not catch status = 1?

That's all it catches. If you are certain your job finished OK (status=1) and yet your Otherwise trigger fired, that sounds like a bug to me.

mandyli wrote:Is this will impact any performance?

Yes, as the documentation specifically notes.

Best Practice is to add it to all jobs set to FALSE. This will allow you to override it when needed for debugging purposes on an 'as needed' basis.

LOL

Or... if you are crafting your own SQL that selects the columns, let the stage generate that for you at runtime. No need for any 'custom / user defined' SQL for anything of that nature.

Note that this is not the same as RCP.

You cannot simply take a record that contains both string and packed decimal data in EBCDIC and convert the entire record to ASCII. Any possibility that is what is happening? You need to only convert the string fields, the packed fields will be happily destroyed if any 'conversion' is done on them. ...

mandyli wrote:Is this good idea to enable '$APT_DISABLE_COMBINATION = True' at project level instead of job level?

No.

Just curious... did you do an exact search here for that error message? There are 12 other posts that include it, many of them resolved so it seems like your solution should be out there.

Well, one question answered. What data type is this field in Oracle? A VARRAY? Post the actual column definition, please. If it is, you can't simply 'drag it to the target', you'll need to process it using PL/SQL via (one suggestion) a stored function so you can 'select' the summed value from the fi...

It may be in an array? Or is? :? Also curious what actual database this is and the internal data type. I would imagine that it (the actual database) provides functions for dealing with this 'array' but hard to say without the requested details. Is "this conversation" actually part of your ...