DSXchange

Sounds like a home grown before-job routine to me. And I've seen people post / use a hack to allow that (or at least what I assume "that" to be) to work but it's never been officially sanctioned or supported.

He explained that. If there's a part in there that is unclear, let us know what.

That is a good approach but make sure people understand the concurrency issues that it brings. You would need to make sure any job that does something like that is always the only job loading that table during that time.

Well, while it may not be a "good idea" I've seen plenty of places where it would not work without them. And if you're worried about exposing them on the command line, then check out the -file option for dsjob.

Is someone telling you to do this? If so, they are messing with you. And as noted, if you really need to store a variable length string with no padding characters, your target database field will need to be changed to a Varchar.

Of course not, that's how a CHAR works after all, being a fixed-width field.

That's because your quotes are being stripped, something DataStage is notorious for doing. A search here for escape should turn up all of the advice that's been posted.

Those are "local message handlers", they need to be in the directory for the job they are associated with and thus no "put in one giant file" option. You'd need to find the job number of the job in question and then migrate the msh file from there to whatever job number the job h...

If this is really what you want to do, use a Routine Activity (where you attach to and run the job using the BASIC API) or Execute Command (with 'dsjob' and no '-wait' or '-jobstatus') stage instead of the JA stage inside your loop. Either can be used to start a job without that pesky 'and then wait...

I'm unclear what step in a job's process is the "loading environment variables" step. What exactly do you mean / are you seeing? Are you saying that the last entry you see in the job's log is the second entry that lists the environment variables from the run? Or only the first one and not ...

So, the error posted comes from the attempt to open the job and then you say you cannot delete it with no explanation as to why. Why can't you delete it? What happens, what error(s) are thrown when you attempt to? The same?

You answer is buried in the above reply - the original post shows everything being passed as a single string, hence no parameter substitution. Break it up as shown to give them a chance to sub into the passed (final) string.

Upsert is never 'more efficient' unless, perhaps, your mechanism to determine one from the other is highly inefficient.

Hmmm... the Job Activity stage waits, period. If you need to start the job and not wait, you'll need to use something else. Question though, if you do that how will you know what happens with all the jobs that you run? Your Sequence job will just fly around the loop and be done, are you planning on ...

Oracle. And unbounded NUMBER field (one with no precision specified) will be imported as Decimal(38,10) since an unbounded NUMBER is stored as a float.