DSXchange

hi..
can any one tell me..that..
if i give one template ...can i make job automatically...generated..
like if i pass source table details and target table details ...
can i do...
automatic job generation...
contact me......
my id phd21@rediffmail.com
phone no: 919886135740

They are trying to get there in Ds7.5 but I think they are still a long ways away from that kind of automation. I think if you wrote one job and exported it and looked at the DSX file then you might be able to reverse enginineer it and generate your own DSX files from your source and target metadata.

Yes, it can be done. A colleage of mine did exactly that, but having worked for Ascential for some time gave him a leg up on someone like you or me trying to work this out on our own. It is not all that simple and not something that someone can just "tell you", but it can be done.

But as Kim said, build your template job and then export and study it. Determine the sections you'd need to generate and work out a methodology to do that. It may be a little easier to tackle the problem using an XML export rather than a DSX, but that would be for you to determine.

Good luck!

him121 wrote:hi..
can any one tell me..that..
if i give one template ...can i make job automatically...generated..
like if i pass source table details and target table details ...
can i do...
automatic job generation...
contact me......
my id phd21@rediffmail.com
phone no: 919886135740

Hi him121,
what you are asking for is extremely difficult in DataStage. I know because I tried for a long time to figure out how to do it.

I implement Sybase IWS and often times we use DS. About 18 months ago I decided having ETL templates for loading IWS would be a good idea and I tried doing this in DS (and INFA).

What I tried to do was create a template for each style of processing and then generate a utility to take the template and generate the new job.

After I figured out what was necessary to make datastage able to generate a new job based off a template I realised it would take well over a month to develop such a utility (if I could get it to work) so I stopped work on it. I left it as just a suite of templates that could be adjusted by hand for each install. No-one paid money for them and they are now with Syabse as 'extra-ware'.

Also, even if I could do this, I felt there was no benefit for me.

How could I get someone to 'pay for a productivity tool for my productivity tool'? No IT manager would give me money for that.

I even toyed with the idea of making it a web service. But now matter how I looked at it there was no money in beign able to generate jobs for DataStage for me.

If the tool was popular Ascential could just build one the same and offer it in the 'next release'.

The basic issue is that DS (and INFA) store the processing requirements with the data model and at a field level. This is required to provide the metadata linkage from source to target which is valuable information. It has the disadvantage that each job records a lot of information.

This is one of the things I avoided with the ETL software I wrote....and of course one of the disadvantages of my software is very little metadata and documentation must be kept in spreadsheets/word rather than the tool....you can't have it both ways....yet...

Isn't this already in version 7.1? The Intelligent Assistant for a migration job runs a user through a wizard where they select the source table and the target table and it builds the job for them. Maybe you can insert your template job into this process.

Can we build our own intelligent assistants? Hopefully Ascential have made this function open source so developers can start building, promoting and swapping intelligent assistants.