A current thread in IBM-MAIN seeks a way to run z/OS DFSORT with a combination of VSAM and PS data sets.
This strikes me as a natural for Pipelines: A collection of SPECS stages each converting a different input format to a common output, all feeding a FANIN whose output connects to SORTIN of CMS DFSORT. But is CMS DFSORT sufficiently Pipelines-friendly that this can be done without resorting to an intermediate temporary file? On z/OS I could do it by ALLOCating SORTIN to the output of a network of POSIX pipes. (Think Samuel Johnson's dog.) -- gil
