DSXchange

You can use an after-stage or after-job subroutine to delete any file that has zero length. You can not prevent file creation by a Sequential File stage.

At least data and stack should be set to "unlimited" for parallel jobs.

You would be better off checking IBM's website. The class is offered in different parts of the world at different times; I do not know where you are, since you did not provide this information in your profile, and I would hate to provide wrong or useless information. The IBM site will be up to date ...

No, you want to keep going and try the other lookups. Maybe Drop on the third.

Yes, Distributed hashed file.

100M ethernet will be a bottleneck. Get a dedicated network segment, get higher speed components (even fibre optic). NICs for fibre aren't expensive these days.

No, they're not. There are no Server SDK routines or Transforms with dots in their names. This is more like the typical PX function naming style, from what little I've seen, hence the question. Not if transcribed inaccurately by the OP. The PX functions also don't have dots in their names; the Modi...

kumar_s wrote:Ray, it seems here most of your post are posted from airport lounge. Is it everyday you fly to your workplace

Most weekends, and occasionally midweek. Highly peripatetic.

What do you need to do with all the intervening lines? Do they become separate columns in a single row? What is the terminating condition for a row? Does each $POUND introduce a new row in the output? Some extra specification would be useful in trying to find a solution: what should be the form of t...

More precisely, your main input stream can contain very many rows too, but only a small number distinct lookup key values.

The idea of a sparse lookup is to be able to avoid loading a huge reference table into either a virtual Data Set or Lookup File Set.

A strange solution to the OP's problem if a fixed number of invocation IDs is used might be to partition the log table based on invocation ID. This is not possible with the current engine (which can only partition on key values, and the log table has a sequential integer key) but might be possible i...

Data with embedded navigation instructions. Rather old-fashioned technology, but it rang some old bells (used to see this kind of thing with paper tape "files"). Tricky to do with DataStage, which is about flows, and can only really remember one row back. You'd need at least two passes through the d...

My understanding is that Trim(Arg) trims "white space" characters, which includes tabs. Next time I'm near a DataStage BASIC manual (as opposed to sitting in a airport lounge) I'll check.

Yes, Abdul is not a particularly Scottish-sounding name!

You might get some more joy from SEMAPHORE.STATUS, but no guarantees.