Runtime column propagation

InfoSphere DataStage Balanced Optimization turns off runtime column propagation at the input link to any target database connector when processing is pushed into that connector, or when the connector is changed to use bulk staging. A DETAIL message is posted in the optimization log when runtime column propagation is turned off. There is no way to anticipate what columns might be propagated at run time, so such columns cannot be included in generated SQL statements. This inability means that, if the root job had runtime column propagation on from source to target in the job, then additional columns propagated from data sources will not be created or populated in the target table.

Runtime column propagation remains unchanged for the parts of the job not pushed into a target connector. When processing is pushed into a connector used as a source or reference, it is possible that additional columns that might have been returned from the original source query will not be propagated from the source database. This lack is because the processing pushed into the source SQL might mask those additional columns that were not named explicitly in the job.