DSXchange

If you're going to do a lot of work in PX do take the trouble to learn C or C++. They are radically different languages from DataStage BASIC; for a start, the C languages are very strict about data types and variable declaration; BASIC doesn't get concerned about either. BASIC does not use pointers;...

Dump it to disk (text file) in an intermediate stage, then read it from there. At the end of dumping any SELECT locks can be released.

And the "%" means add leading zeroes as a mask. Another way to write this would be to use zero as a background character. Fmt(YR,"2'0'R") : "-" : Fmt(MO,"2'0'R") : "-" : Fmt(DA, "2'0'R") The zero has to be quoted to differenti...

Take a second output from each Job Activity with a "Failed" trigger.

All of these go to a single "Any" sequencer.

The output of the sequencer runs into a single Notification Activity that can report that "something" has failed, but not more exactly than that.

Using a Hashed File stage you can only perform a key-based lookup.
You can perform a reverse lookup against a hashed file stage using a UV stage, but will need some tricky SQL to handle the possibility of NULL (lots of OR conjunctions). And it won't be fast unless you index the non-key columns.

If you really want to do it in DataStage you can, though I'd fully support your using Quality Stage - after all, "bucketing" address components is one of the things it does supremely well. For those who don't have Quality Stage... A "final delimited field" function is handy thing to have in one's to...

Pivot stage is exactly the right tool for your stated requirement.
The examples in the manual (pivot.pdf) will show you exactly how to achieve it.

If you got the DSX file by exporting, all the components in it are probably still in your project. When you use the Import menu in Manager, choose the "import selected" option, so that you can see what's in the DSX file prior to importing those components (and allowing you to select some, rather tha...

That's known as the Old MacDonald's Farm error ([EIO] I/O error) Yes, it looks like at best file corruption, at worst hardware failure. The subroutine in RT_BP415/JOB.1433470019.DT.1330157629.TRANS1 - which is the code generated by a Transformer stage - should give you a clue as to which file was be...

Use link variables on the Output link. However these are not available until the Output link has executed its "Put" operation, so you'll need a second output link (let's say called Output2) that is executed after your Output link. db2/udb | | <lookup> | -----> Transformer --> Output | | V Output2 On...

Jim, would you be prepared to share that style sheet? I was about to embark on a similar exercise! (You can email me via my profile.)

DRS is used with the Peoplesoft jobs because not all Peoplesoft customers use Oracle as their database. Indeed the DRS was specifically developed for this kind of deployment, so that the database type could be made a job parameter. The reason you should prefer DRS to the Oracle stage type is insuran...

MKS Toolkit spoils one sometimes! How about something like this before/after subroutine? SUBROUTINE RemoveFinalTerminator(InputArg, ErrorCode) ErrorCode = 0 OpenSeq InputArg To hFile Then * Position to a point two characters short of EOF Seek hFile,2,-2 Then * Read those two characters to m...

Without using a stage type that supports BASIC expressions (such as the "BASIC Transformer" stage), the parallel extender does not support the concept of an internal format for dates. You could, of course, create a buildop that uses substringing techniques to decompose the dates into "days since a c...

You really should have either posted this on the Server forum or indicate that it's a parallel job. Which engine it is will govern the answer.