DSXchange

Hi folks

I have a strange problem. Here goes.

We all know DataStage is strictly case-sensitive.

I thus have 2 environment variables, both pointing to the path where my .ds (dataset descriptor) files are stored - one if PathDS and the other is PathDs. this is because at some point we were using DS and eventually norms changed and we moved to the naming convention of having just the first letter in caps thus converting to Ds. No one thought about removing the PathDS from the list of environment variables.

Today, in my job I can addboth these variables in my parameter list (job properties->add environment variable) but cannot compile since I get the message Precompile failed. Found duplicate environment variable in job definition.

Is this funny? How can I add duplicate variables without any error but am only policed when i try and compile the job?

And it gets better:

If I don't declare the variable with the small s so PathDs but use it in the job, the job runs fine if launched unitarily but fails when called by a parent sequence job. However a view data on the stage does not run. This fails with the error that the dataset inidicated by PathDs/dataset does not exist - which is logical to me.

What is flabbergasting is how is the job running?

Any ideas?

Thanks

Tony

Forgive me... it's 5:30 in the morning, the coffee isn't ready yet and I haven't had much sleep... but I think what you'll find is that it isn't DataStage that is case sensitive, just UNIX.

I was under the impression that both DataStage and UNIX are super case-sensitive - at leats that's what Ascential had taught us in 2005...:-O

Well.. we know certain aspects of DataStage are case-sensitive, like parameter names for example. I was just thinking that in your particular scenario that there was a mismatch between sensitivities, so to speak between the tool and the O/S level. But then in my defense it was very early in the morning and I haven't touched the tool since the end of 2009 so... perhaps someone else will step up and enlighten us both.

TonyInFrance wrote:How can I add duplicate variables without any error but am only policed when i try and compile the job?

Precompilation happens on the client tier, and in Windows, environment variable names are not case-sensitive. The rest of compilation takes place on the DataStage server.

Ah... so there is the mismatch of which I spoke. Mostly.