Am 22.08.2015 um 12:19 schrieb Melinda Varian: > The end of that discussion in July of 1999 was this REXX stage by Michael > Faulhaber: > [...]
Two remarks: i - it was not my intention to end that discussion in 1999, and ii - even the CSVCLN REXX shows '.....Mike' as author it is not at all "my" program. It is an assembly of ideas I found here and there and help from this list -- see the comment about the 2x speed up due to a hint from Melinda. The result I have put in a template (skeleton stage) from Steve Hayes. Why do I insist that it is not my routine? Short answer: I am not allowed to publish my programs I did at work. Long answer: independent from this thread few month ago I asked my chief about a permission in written to publish old programs, at least since many doleful years with no VM we may not run them any more. As I may not make money with it he agreed alas only by word of mouth. I got warned from the workers council, too unsafe in those times when 50% of the jobs are axed. I feel sad about this situation, I got so much help from this list but I may not show my solutions. I am very sorry. My only consolation is, with my approach to Pipe I could deliver John a reproducible way to a very seldom error he tried to nail down since long, he said. :) Back to the subject: CSV is used for long time archiving data in documentation institutions for its simplicity, with little effort it may be exploited also on future platforms. It was fun using Pipe for CSV files with success where Excel failed on the PCs of 15..18 years ago. What I found useful: i - everything is text, if there are numbers they are represented as text string of numerals. This ensures save data transfer without failing automatisms when reading it with some intelligent software. ii - use column headers made of single words. This helps for setting tabs in XEDIT before a COMP *. Set the tabs with a TABS prefix command that does: > "COMMAND LOCATE :" !! wo 'COMMAND EXTRACT /CURL/' > do i = 1 to words(curline.3); z = z wordindex(curline.3, i); end > "COMMAND SET TABS" z (this is an excerpt from a XEDIT file containing all prefix commands from Uni Konstanz, Germany). Using column headers in your routines makes them a bit more complicated but gives you the freedom to change column sequence in the input file with no need to change your routine. iii - For testing ideas use COL XEDIT (VMTOOL from Gary Vair) on small, expanded tables then code it as pipeline for a CSV table. Many things I did without programming, just PIPE in FILELIST or directly on CMS was enough, only for complex and recurring tasks I made programs. Ciao.....Mike -- www.Ok.de - die kostenlose E-Mail Adresse
