Dynamic MetaData

[suman27]

Hi..
Iam Pretty new to DataStage. Iam trying to build a re-usable job for my project.
Is there any way to papulate meta data at runtime, so that I can use the same job for different sources having different metadata.


Regards,
Suman.

[ray.wurlod]

Yes, but probably no. It depends on how dynamic you want to be, and how your job design needs to affect the data.

A combination of schema files and runtime column propagation can mean that reasonably unspecified data can flow through but, as soon as you need to apply any kind of transformation, your design must name the column(s) to be transformed. Unless that column appears in all of your "dynamic" schemas, there goes your idea!

[vmcburney]

You want dynamic metadata when there is minimal column renaming and data transformation happening in your job. So it's more a type of data replication than real ETL. DataStage jobs are easy to build when you load and define your metadata at design time, they become trickier to write, but re-usable, when you try to manage your metadata in run time schema files, SQL files, SQL job parameters etc.