On Mon, 6 Feb 2012 15:02:42 +0000, Martin Packer wrote: >Looks to me like it's being greedy. Look up the term "greedy" in relation >to Regexps and you'll see the match is much wider than you anticipated - >matching many more characters. Not sure if awk can do non-greedy matching. >But there are usually workarounds if not. > Mostly, it can't.
>From: Ken MacKenzie >To: [email protected], >Date: 06/02/2012 14:53 > >I'm not sure if this is the appropriate forum, please point me to the >correct one if it's not. > There's also MVS-OE. >I'm playing around with regular expressions and I want to achieve the >following. I spoke to a Unix geek but he didn't really understand what I >was asking. > >Given the following sample data, I want discover only the first occurrence >of any string which matches my regexp. >QQQQABCDEFGNOPQRXXXPPPPABCDEFGNOPQRYYYOOOOABCDEFGNOPQRZZZ > >I tried: awk 'sub(/CD.*QR/,"junkt")' fxdata in an attempt to change >QQQQABCDEFGNOPQRXXX to QQQQABjunktXXX but instead, it takes the final >occurrence of QR, and returns QQQQABjunktZZZ. Notice the ZZZ on the end >instead of XXX. > You might try: sub(/CD[^Q]*QR/,"junkt") # Wildcard matches no string containing "Q". >This is being driven from a REXX exec in ISPF, if any of the above is not >clear, I will try to explain further. > Eek! Are you using SYSCALL spawn in an EDIT macro for that? I thought only John M., Kirk, and I did that sort of thing. --gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
