Runtime Column Propagation (RCP)

[skarthikcs]

Hi,

What is the purpose of RCP in paralell jobs?
Does checking or unchecking of this option really make any difference in performance of the job ?

[SAMEENA]

Hi,
RCP--Runtime Column Propagation is mainly used in the scenarios where the modification in DDLs is possible.
The advantage of having RCP checked is - if there is an addition of columns in your source table and these columns need to be propagated as such to the target (without any derivation) then you just need to add these columns in the source meta data and target meta data and for rest all intermediate stages you can simply check the RCP option.i.e. you dont need to add these columns in the intermediate stages.
And if the source is a sequential file then you just need to modify the schema file and you don't even need to add the columns in source meta data.

[ray.wurlod]

On the down side, lineage analyses will not be reliable if you use RCP - the columns will appear to have magically appeared out of thin air.

[manuel.gomez]

Sorry if the question sounds stupid, but I am so new to Parallel Jobs....

Does RCP have any sense when you only have one intermediate stage between source and target stages? For example:

Seq File -----> Modify Stage ------> Seq File 2

Thanks a lot!!!