DataStage Version Control

[ray.wurlod]

The first and obvious advantage of Version Control is that it's free. It ships on the DataStage CD (versions 5.2 and later).
I have found that it's straightforward to use, but lacks strict check-in/check-out capability; one needs to impose a layer of management (rules for developers). One thing it can do is make everything read-only in the test and production environments. However, there is nothing in Version Control to prevent two developers working on different aspects of the same set of jobs.
Other version control software, such as PVCS or VSS, do require a substantial amount of extra effort if you are going to use it to control all DataStage objects, because the definitions of DataStage objects are stored in hashed files in the Respository, and the software has no knowledge of their internal structure.

[WoMaWil]

3 years ago, when we started with DataStage there was no Version Control available an though we also had several shell-scripts and sql-scripts in our project we decided to take PVCS. For the complicated manual process transfering jobs, routines and transforms from and to DataStage we have written a client-commandline-tool which we call dsmove.

Its syntax is:

dsmove
/server
/user
/password
/project
/direction E[xport]|I[mport]|C[ompile]
/type J[ob]|R[outine]|T[ransform]
/name |||
[/path |]
[/overwrite]
[/compile]
[/append]
[/delete]
[/logfile [path>]filename>]
[/checkpwd]

since then (version 3.6) until now (version 5.2) it servs well our needs.

Wolfgang

PS: The company where I am employed to does sell this tool to those who like to have it.