DSXchange

You need a key. If nothing qualifies, make one up! For your situation where you have a single row, hard-code the key to something like a "1" and then use that (again, hard-coded) in your job. Works great!

From what I recall, DataStage is stripping part of your command out before it actually executes it... the single quotes, I believe. You'd be much better off if you wrote a script that does the commands you are executing and instead use DSExecute to simply run the script, passing it the parameters yo...

That's what I've been wondering as well, but knowing absolutely nothing about the specifics of AS400-ness was unwilling to take point on that particular trail. Thanks, t!

Well... let me throw something out on the table. Something I setup way back in version 3 dot something or other when I wasn't strong in Universe Magic. To Peter's original question, how to tell when polling for something that a number other than zero have been found. I setup something similar to wha...

More than likely - if you are sure you are correctly populating the hash file with the current business keys in the table - you are facing a trim problem. Perhaps the key field has trailing spaces and your csv data does not, or vice versa? That would cause every lookup to fail. Also, when you insert...

Doh! I meant to type ADN, as in the Ascential Developer Net. Somehow, DSN came out instead.

It means at least one of your column names is longer than 30 characters. Since it can't be in your table, then somehow it made it into your job that way.

Do they have the same metadata? If so, you can do a couple of things. Check out the Merge stage, it will allow you to combine the two files as if they were tables. Combine the two via an O/S command before job so that you only have one file to process. If they are different layouts, perhaps one coul...

If it works anything like a Server job lookup, then you have two options: Don't check the data fields in the lookup, check one of the key fields for NULL. If you put a value in the key field and it comes back NULL, then you know the lookup failed. Is there a 'status' you can check? In Server jobs yo...

Wrap it in a Sequencer job. That would greatly simplify how you could send an email for any problems in your job. If you specifically want to only email when the problem is a 'database down' type problem, one 'trick' is to first use a Job Activity stage that only does a Validate on the job. If it fa...

Well, here's the party line... As long as both links come from the same Transformer, you should see a new option on Transaction tab. In addition to Transaction Size, there should be one for Transaction Grouping. There you indicate what action you want to happen when one or the other of the actions f...

Have you tried running this command from outside of DataStage? Suggest you get it to work first from the command line, then transfer it into the routine. You should also ditch the semi-colon before the pipe.

You don't need a Link Collector. There is a FAQ posting here which may help. You would split the record up using the substring operator, add the terminators and write it out as a single "record". When you read the output back in (later), each input record has been turned into three output records be...

I would think you could use the Sequential stage, tell it the file is pipe delimited and bring the first four fields in as one composite field. Then, in a transformer, break it up properly into four fields using the FIELD function.

Depending on the nature of what you are doing, they are trying to address situations like this. If you have 7.5 there are new Sequencer stages to control "looping" within a Sequencer job and from what I've read the LoopStart stage has some ability to pull information from a delimited source. I expec...