I wanted to check, if I want to develop a job which enriched with parallelization. Suppose, job contains, sort, tMap used for join 1,2 files/table, other component. So I wanted to understand, how this will work?
what is best way to build the job, so performance will increase and in addition, output with and without parallel is same.
@manodwhb, I can use store on disc option when data is huge or not fit in to the memory.
My question is different. My question is about implementation of parallelization. and How it work when we have component sort, tMap and etc in the job. what is run time behavior when job set to run in parallel consisting of these component.
@manodwhd, I know how to enable the parallelization, but I wanted to understand how it behave at run time. likeif I join 3-4 files to tmap, then how it will join with main flow which is parallel and dedup component or with other components present in job.
