Read a dsx file

eoyylo · Post by **eoyylo** » Tue Oct 14, 2003 10:54 am

Hi,
i'd like to read all information contained in a dsx file (an export of a DataStage's project).
I created the file with the only option "Job define" in the export tool.
In the file there are a lot of reserved words. Can anyone tell me where is possible to find documentation about them and documentation about the file's structure?

thanks in advance

Mario

kduke · Post by **kduke** » Tue Oct 14, 2003 12:13 pm

Mario

This is not documented anywhere. It is an Ascential file. If you want to process this same information then export in XML format. It is the same information in a standard format. Both of these are basically dumps of the DS_JOBOBJECTS hash file.

Kim.

Amos.Rosmarin · Post by **Amos.Rosmarin** » Tue Oct 14, 2003 3:01 pm

Hi Mario,

If you'll specify the reason you need the dsx maybe you could get a better answer.

Looking at the DSX file you'll see that it's stracture is like xml - It has a fixed header and each element has an opening and a closing tag . Compare the dsx to the job you can link the different tags to satges, links, table definitions and the rest of the objects on the canvas.

HTH,
Amos

ray.wurlod · Post by **ray.wurlod** » Tue Oct 14, 2003 5:10 pm

You can even do two exports - one in DSX format and the other in XML format - and compare them.

eoyylo · Post by **eoyylo** » Fri Oct 17, 2003 9:39 am

Hi,
I want to read the dsx file (or XML, is the same) because i'd like to write a program that
1) can determine the resources used (file, oracle table etc) in order to calculate the execution precedences
2) extract the oracle query
3)plot the data traffic

i can't use metastage because the job's are developed in several server with different tnsnames.

Now i have hundred of jobsand it is impossible to calculate by hand.

I have understood the greater part of the reserved words but someone is mysterious.

Mario

kduke · Post by **kduke** » Fri Oct 17, 2003 9:53 am

Mario

Why not use the Reporting Assistant (doc_tool.mdb)? It is in the DataStage Manager. It extracts all the SQL. You can connect to different DataStage servers. It is a little difficult to find it but we can help you.

When you say data traffic, what do you mean? I think most of us need to calculate bytes moving through each job. I think that would be a benefitial number to understand performance.

Kim.

ray.wurlod · Post by **ray.wurlod** » Fri Oct 17, 2003 8:08 pm

The reporting assistant (documentation tool) is to be preferred for another reason. If your data source names, user IDs, passwords, pathnames and the like are parameterized, which is recommended best practice as they can differ between development and production systems and/or can change over time, then all you have in the DSX file is the job parameter reference, for example #DSN#.

In the documentation tool you can create a join to the JobParameters table and the Properties table to decode at least the default values for these parameters.

To get the run-time values you need to have captured the process metadata as well. This information is in the job log (in the "Job Started" event). Note that encrypted parameters are reported using a standard string, so as not to give away, or even hint at, the encrypted value.