converting shell scripts to datastage jobs

[rajiivnb]

Hi,
My project requirement statement is as follows:
1.Existing project is in production with shell scripts and datastage job together,where the shell scripts talks with some complex queries and as initial process before the datastage jobs.Scripts written in unix talks more with the extraction phase.Production is happening with long time taken to complete the process
2.this exisiting shell scripts has to be converted to a datastage jobs such a way that time taken to be considerably less when compared to shell scripts .

Please someone suggest how will i go about initialize this process

Regards,
Raji

[ray.wurlod]

There is only one sensible way to go about this task.

1. Analyse what the shell script does. Under all circumstances.

2. Write this out as an set of target from source mappings.

3. Implement the same as one or more DataStage jobs.

Look for opportunities to optimize at each step in the above.

Since you've specified parallel as your job type, you might also examine opportunities to implement partition parallelism (see Parallel Job Developer's Guide) either through the PX engine's inbuild capabilities, or using multi-instance server jobs each acting on a subset of the data to be processed.