DSXchange

Always was going to be DataStage 8. It's just that the double-secret hush-hush internal project name for it was Hawk... and then someone blabbed.

Yuck... that will take for-ev-a. Look into making it a two step process... first extract and transform, landing it to a flat file. This job can be 'multi-instanced' rather easily, using mod in Oracle and a 'partition number' as the Invocation ID. Make sure the output files have a unique name as well...

Not sure what else to say other than you generate it and pass it into the job as a parameter. The 'old fashioned' way would be to create a routine that gets a system timestamp and strips out all the yucky bits you don't want, typically the dashes and colons. From '2006-11-28 20:44:13' to '20061128_2...

More than likely you are wrong there and they in fact do NOT match. That is the most common problem when lookups 'fail' which is why Shane ask about it first. Your other problem could be your lookup sql - is it generated or user defined? Do your keys match up properly with the positional parameter m...

Or pass in the timestamp as a job parameter and concatenate the two pieces in the FTP stage.

I understand and sympathize... however, this isn't really a 'DataStage' issue. When you sign a license agreement with pretty much anyone, they prohibit you from distrubuting that software out to any other parties - and that typically includes training materials as well. Not that you won't find peopl...

samsuf2002 wrote:how can we set time out in data stage ?

Simple answer - you can't.

Certainly you can ensure that the proper status is returned by any shell script that DataStage leverages. The question is how to do it when those errors are buried in a call to another tool, I assume? We have something similar in the Oracle world, scripted calls to sqlplus that can generate all kind...

Who says you can't?

Leave the Timeout Length set to zero and you get a single check.

He's not saying that my Guru friend. It is exactly what you are saying - process X files and in-between every iteration look for a file that says you should stop any further processing. Break out of the loop rather than going on to the next file...

That's not bold, that's just stupid. Sounds like the FUD that Core Integration spread here regarding anything DataStage - IBM, Hawk, EE, upgrades, everything was fair game and all kinds of nonsense and out-right lies were spread... whatever it took to sell Informatica. Problem is it worked because t...

Answer the question re: the Oracle field - DATE or TIMESTAMP?

If you hack off the microseconds like you've just shown then it will load into either field, however you are doing yourself a disservice if you do that and load it into a TIMESTAMP field.

As noted, the Sequence job only knows which activity to restart with. If you need a job to restart internally from 'the point where it left off' you need to build that functionality into the job itself.

I know how I've done it in Server jobs, PX jobs on the other hand... [shrug]

Not only have you replied to a three year old thread but you've asked for something that is technically illegal for someone to provide you.

What exactly is the problem you are waiting on a resolution for? With cryptic messages and obviously edited posts from other folks, I can't really tell at this point.