Page 1 of 1

Re: Migration from Abinitio to Datastage

Posted: Fri Aug 11, 2006 10:10 am
by shawn_ramsey
There is the idea of runtime column propagation that does the same thing that you are describing.

Re: Migration from Abinitio to Datastage

Posted: Fri Aug 11, 2006 10:29 am
by psriva
Do you mean that it automatically maps input columns to output columns ?

Thank you.

Re: Migration from Abinitio to Datastage

Posted: Fri Aug 11, 2006 1:11 pm
by jking123
In Enterprise edition the two features that come close to the dynamic metadata supported by abinito are schema files and runtime column propogation. There are quite a few restrictions to this though e.g. it is one way i.e. you can't use schema files like you can use dml files on the output side. RCP is also very very sensitive to data types i.e. it is just not column names that need to match but also types and that too exactly.

There are lots of other differences.

Posted: Fri Aug 11, 2006 1:59 pm
by Jay
There is an option in the transformer where in you can select "automatically map columns" having the same name. Other than that i dont know anything similar to the in.* you are asking about.

Posted: Fri Aug 11, 2006 4:09 pm
by kduke
I was told IBM/Ascential had a tool to automate this conversion as well as converting from Informatica. It does not convert everything but supposed to convert most. Ask them. Ray may know.

Posted: Fri Aug 11, 2006 4:53 pm
by chulett
Interesting... I know they have one for Informatica but hadn't heard of one for Abinitio. Curious what comes of that investigation.

Posted: Sat Aug 12, 2006 4:06 pm
by vmcburney
I wouldn't be surprised if there was a single tool that had a conversion engine and had various adaptors for Informatica, DataStage server jobs and Ab Initio. In theory they could build a converter for any ETL tool if the demand was there.

Posted: Sat Aug 12, 2006 10:28 pm
by ray.wurlod
I believe the Ascential solution (?) was to throw a heap of (offshore?) resources at it.

Posted: Tue Aug 15, 2006 9:09 am
by psriva
For the functionality of in.* of Abinitio to be placed in Datastage to make the columns propagartion easy I have enabled the RCP 3 ways.
Once in the Administrator, and once within the Job Proprties and once within the stages.
When I do this, the columns get mapped by themselves and I dont have to do anything in the transformer.

Thanks to everyone.

Ab Initio to DS Convertion

Posted: Tue Oct 23, 2007 11:47 am
by rajeev_prabhuat
kduke wrote:I was told IBM/Ascential had a tool to automate this conversion as well as converting from Informatica. It does not convert everything but supposed to convert most. Ask them. Ray may know.
Hi Duke/Ray,

Do we have such tool for convertion from one tool to another bulit by IBM?

Regards
Rajeev Prabhu

Posted: Tue Oct 23, 2007 12:16 pm
by chulett
If such a thing does exist, they won't simply give it to you. In my experience, it would only be offered as part of a consulting package.

Posted: Tue Oct 23, 2007 3:50 pm
by ray.wurlod
Indeed, such tools are jealouly guarded by IBM, and only available as services (that is, you never get to see the tools and they're never deployed to your hardware).

As to whether they've developed one for Ab Initio, I have no idea. Why not ask IBM, specifically the Center of Excellence for Data Integration (CEDI)?

Posted: Tue Oct 23, 2007 3:54 pm
by trokosz
The "tool" is called DS Migrator and is managed by IBM CEDI Professional Services as an offering. Much of the migration must take place overseas due to legal issues with Informatica. The tool addresses converting Informatica Maps (or even DS Server to DS Parallel)....This was all initiated when DS replaced Informatics in the Peoplesoft EPM ETL offering and Peoplesoft wanted some automated way. AbInitio is not part of this conversion offering.

Posted: Tue Oct 23, 2007 5:28 pm
by vmcburney
I sat in on an IBM CEDI presentation last week at IOD on re-usable DataStage parallel job flows. They used DataStage jobs with column propagation, parallel stages such as modify and sort that received instructions via job parameters with all the metadata defined in external schema files. In one project they spent eight weeks building four re-usable DataStage jobs that could handle hundreds of different interfaces. The tricky part is managing all the schema files as they cannot be created through import wizards. Most DataStage projects go for hard coded column metadata because it's easier to manage.