DSXchange

Multiple Value properties, as suggested by balajisr, is the way to get this to work. I've done lots of them.

Maybe. Probably. But not if the two flows are occurring simultaneously and writing to the same sequential file. This is an operating system restriction, not a DataStage restriction. One writer, or N readers. And you may run into locking (deadlock) issues if you are trying to insert the same keyvalue...

Define "too long".
Surely it takes as long as it takes (all else being equal).
If it takes two hours and you expected it to finish in one, change your ideas so that you expect it to finish in two.

Why do you want to do this? DataStage's main strength - the way it's always been marketed - is the amount of coding/scripting that you DON'T have to do. You CAN delete a job from within the GUI. That's the DataStage way. Your assertion, therefore, is totally incorrect. As it happens, one CAN delete ...

Load values from row #1 and row #2 (as detected by @INROWNUM) into stage variables and from there into output columns. Constrain the output @INROWNUM = 3.

Multiple users are managed by the operating system, whether UNIX or Windows. The same considerations apply to use of groups and DataStage roles. The only difference is how you set up users and groups; in Windows there are GUI front ends (Windows (!)) for doing these tasks.

One way is to change your expectations.

The simple answer is that there are no data types in server jobs (internally), but parallel jobs executed in a strongly-typed C-language environment.

We have no idea what your job design is, so it is impossible to recommend anything apart, perhaps, considering using multiple parallel streams of processing, if your operating system has sufficient spare resources.

Was able to see it again today. Mind you, I'm not using my usual machine; I'm using an Arabic Windows 2003 Server with IE6. There's a workaround - refreshing the page updates the icons to "all have been read" status. But without explicitly refreshing the page that does not occur. Will continue to mo...

That particular one doesn't seem to be private.

NEVER kill DataStage processes.

Is it me, or are the read/unread icons on the index page not resetting to unread once all the posts in a forum have been read? This appears to be new behaviour since yesterday.

How do you propose to guarantee the "+1" requirement across an arbitrary number of parallel processing nodes? Are you going to introduce a bottleneck by forcing a Column Generator stage to execute in sequential mode? Is this, indeed, a sensible requirement in a parallel job?