DSXchange

Yeah , I have observed this behaviour as the ASCL treats '' as @NULL so it goes on inserting NULL (only WRT to SQL). Note that this is not "Ascential" behaviour. DataStage knows the difference between an empty string and a null. However, databases are what generally consider them to be equivalent i...

No, no new setting required. Do you 'source' the dsenv file first before running the dsjob command?

Well, then you need to work through the other stages one by one until you find the culprit.

However, since there's no object code in a Shared Container to fall back on, why in the world would you want it while someone is (supposedly) in the middle of changing it?

Of course, if they are just a tourist...

If you don't properly size a hashed file for the number of rows you will be sending it, it starts to 'overflow' and slows down. When it overflows badly, it starts slowing down badly as well - until the job seems to crawl or stand still. One test would be to replace the hashed file with a Sequential ...

Welcome! :D The FTP stage is kind of a pain to use because it is metadata driven, unlike any other ftp. So you will have 'issues' like this using it for a file where the metadata changes from row to row. Try defining the record as a single varchar field, telling it there are no delimiters and probab...

What stages are you writing to in the job? They would be a typical culprit in a situation like this, especially a hashed file that's not properly sized.

A 'not compiled job' would get you a specific error about it not being in a 'runnable' state. This, from what I recall, is some kind of permissions issue down in the project (or engine?) on your server. Search the forum, as I'm pretty sure the problem and solution have both been posted more than once.

Nothing really, when you get right down to it. Both serve as "job control" with the biggest difference being that fact that DataStage writes all of the code for a Job Sequence while you write it for a Batch job.

No clue, but you might want to post this over at ADN. There's more TX traffic over there as well as a fine gentleman by the name of jsmoley who would love to help you with the MapStage plug-in.

First off, delete your duplicate post so we don't get two disjointed conversations going. Second, after the job aborts try to Reset it in the Director. Sometimes it will pull a record (from &PH&) into the log labelled "From previous run..." and that will have more information about the abort.

ArndW wrote:the column is probably an Oracle VarChar2() datatype ... and by definition the loader trims these and makes it a null field.

There is a sql loader option to PRESERVE BLANKS and I would think there would be a way to enable this for a load.

What is the datatype/sizes of both fields in your job? What exactly happens to negative numbers?

Welcome. :D You can only Add/Delete projects in the Administrator if you are logged in as "dsadm", otherwise they are greyed out. The CPU count is a comparison of your license to what it detects on the server. In this case it means it detected 2 CPUs and that you are licensed for 2 - hence the fact ...

Hi Murali, welcome.

I think you'll find that StringToDecimal is a PX function and not one that can be used (or is even really needed) in a Server job.