DSXchange

There's your first problem right there. From where do you plan to "capture" the values?

edward_m wrote:Thanks..i have 97 nullable columns.So its hard for me to define '*' to each and every column based on their length.
Is there any alternative solution ??

Thanks in Advance..

No.

It's not hard. It's just tedious.

Look at the generated OSH. Unfortunately this IS how it works. Once first wait starts waiting, nothing happens till it returns. You'd have to "roll-your-own" to implement simultaneous waiting. This would not be difficult in a before-job subroutine or in a routine called from a Routine activity. The ...

I seriously doubt you can have a "passive" stage with an input and an output in a parallel job. This would be a blocking operation - which would interfere with pipeline parallelism. These are banned in the parallel job environment - the output link would be treated (and painted) as a reject-handling...

Even inciting to crime is a criminal offence.

Stick it in a directory (which Oracle can access as an "external table" - Larry-speak for a directory).

There are LOTS of processes. The player process PIDs can be logged by setting APT_PM_SHOW_PIDS environment variable.

You could have pre-read input port #1 into a static vector, as was suggested; you can then access this as often as you wish. For better performance, use a hash table rather than a vector - a bit more coding, but probably worth it.

It might also be that the inode table for that particular file system is full, or that the file has reached 2GB and large file support is not enabled. Check with your UNIX administrator.

Start by disabling operator combination, either at the stage level or at the job level ($APT_DISABLE_COMBINATION as a job parameter) to isolate the problemt to the actual stage that is generating it.

You may discover that the problem is not in the Transformer stage, in spite of your assumption.

What part of Surya's code is unclear? It looks like quite an elegant solution.

Does a VOC pointer exist for MyHashFileName ?

It's hashed file, by the way, not hash file.

It's a way of formatting to ensure leading zeroes. Look at the Fmt() function for possible values. Following an expression with a space and a format designator is another way in DataStage BASIC to achieve the same result. I could equivalently have written: Ans = Iconv(Fmt(NewYear,"R%4") : "-" : Fmt(...

Every server job is dynamic - it potentially processes a different number of rows each time it is run.

Plenty of documentation, with the occasional witty aside, is another way to achieve some form of dynamism.