I didn't know about the existence of the RFC. I think that some things are missing, for example: in Germany (and German editions of Office packages), the separation char in CSV is a semicolon, because the comma is used in numeric values instead of the decimal point (see COBOL's "DECIMAL POINT IS COMMA").
So many tools including DB Visualizer allow to specify the separation char when writing CSV files. This should be mentioned in the RFC. The RFC is still valid, even if the separation char is semicolon, or Tab. For the line separation sequence, the Unix variant (LF only) should be allowed, too. Kind regards Bernd Am 17.11.2016 um 16:25 schrieb Bill Woodger:
CSV has been around a long time (I certainly used them in '91, and perhaps earlier). It for sure would be good to have native COBOL support, but 25+ years without it may indicate it is not much of a priority. I'd support an RFE, but not lose any sleep about it happening any time soon. I don't think the CRLF should be done within the program (except the "escaping" has to be considered) as the correct things for the target OS can be added by the "file transfer" mechanism. Otherwise the RFC in question is as expected. There are probably hundreds/thousands of "roll-your-own" COBOL CSV-processors out there, and probably with lots of bugs. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
