Datastage cross references

guidomarcel1 · Post by **guidomarcel1** » Wed Sep 06, 2006 8:08 am

Hi,
I am interested on a tool which allows to help me answer the following questions:
* which job is reading which table / view
* which job is manipulating a certain data element
* which job is updating a table /view

Is there any tool on the market to perform such cross reference analysis?

Many thanks
GuidoMarcel

kduke · Post by **kduke** » Wed Sep 06, 2006 9:03 am

DwNav which I wrote. MetaStage which IBM sells.

ray.wurlod · Post by **ray.wurlod** » Wed Sep 06, 2006 4:50 pm

Usage Analysis in Manager.

All these tools require that your metadata management has been exemplary, that you have not invalidated the linkage between the table definitions in the Repository and those in the jobs.

You can determine (version 7.5 and later) whether these are intact by opening a Columns grid in a stage in a job. Right click in the grid itself and choose Properties. Enable all the hidden fields, particularly the final one (source table definition, or some such name).

guidomarcel1 · Post by **guidomarcel1** » Fri Sep 08, 2006 8:30 am

Thanks for your anwers.
As Datastage is generating BASIC code, I was hoping that there would be a tool which scans the BASIC code. These kind of tools are very popular on the mainframe (good old IBM, COBOL, ...) .

Regards
GuidoMarcel

kumar_s · Post by **kumar_s** » Fri Sep 08, 2006 8:38 am

Each job will create its own BASIC code, but still you have set of hashed files, where it acts as a common repository for the metadata of datastage. Where the guys who answere above in your post were very much expert in extracting the required data from it. Generally it is not suggested.

kduke · Post by **kduke** » Fri Sep 08, 2006 11:14 am

The BASIC code has an engine which controls it. It is a nightmare to extract anything from it. Kumar is correct it is much easier to get metadata about the job designs from the repository tables like DS_JOBOBJECTS.