Thanks for the suggestion. In general, I'd like to stay with NAMES-style entries because all those other tools you mention use them, so IMHO it's more likely for IBM to support it long-term if their own tools use it. I had looked at writing an analogue to the functions of the CMS filesystem (open, read, write, update, point, close) and using Pipe structures to compose/parse the records, but the lookup functions built into NAMEFIND eliminated a lot of code and the format is human-readable. I had also experimented with using the vardump/varload stages to essemtially maintain a big REXX stem, but that got fairly baroque in short order.
Worth a try though, I suppose. On 8/18/20, 7:30 PM, "CMSTSO Pipelines Discussion List on behalf of amphitr...@ok.de" <CMS-PIPELINES@VM.MARIST.EDU on behalf of amphitr...@ok.de> wrote: Sure: CSV.