RegEx question

vit has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: RegEx question by ikegami (Patriarch) on Dec 21, 2017 at 22:14 UTC
If you mean "something that isn't `bank`" (as you said), `/ ^ (?: .{0,3} \| (?!bank}.{4} \| .{5,} ) nova [ ] scotia (?: .{0,3} \| (?!bank}.{4} \| .{5,} ) $ /sx` [download] Alternative: `/ ^ (?: \| [^b].* \| b[^a].* \| ba[^n].* \| ban[^k].* \| bank.+ ) nova [ ] scotia (?: \| [^b].* \| b[^a].* \| ba[^n].* \| ban[^k].* \| bank.+ ) $ /sx` [download] If you mean "something that doesn't contain `bank`", `/ ^ (?: (?!bank). )* nova [ ] scotia (?: (?!bank). )* $ /sx` [download]	[reply] [d/l] [select]
Re^2: RegEx question by vit (Friar) on Dec 21, 2017 at 22:41 UTC
Is it possible to write it without Pattern Modifiers? In other words without /sx	[reply]
Re^3: RegEx question by AnomalousMonk (Archbishop) on Dec 21, 2017 at 22:48 UTC
Sure, just squish it all together (untested) `/^(?:(?!bank).)nova[ ]scotia(?:(?!bank).)$/` and remember that `.` (dot) no longer "matches all." But why would you want to? Update: Actually, you no longer need the character class (update: because without `/x` a space is a literal space): `/^(?:(?!bank).)nova scotia(?:(?!bank).)$/` and if you just gotta get rid of that last space: `/^(?:(?!bank).)nova\x20scotia(?:(?!bank).)$/` (both still untested). Give a man a fish: `<%-{-{-{-<`	[reply] [d/l] [select]
Re^4: RegEx question by vit (Friar) on Dec 21, 2017 at 23:10 UTC
Re: RegEx question by Laurent_R (Canon) on Dec 21, 2017 at 23:52 UTC
It might be simpler two use two regexes: `# if the string is in $_: /nova scotia/ and not /bank/; # else: $str =~ /nova scotia/ and $str !~ /bank/;` [download] And that second option should also work in Java with the appropriate syntax adaptations.	[reply] [d/l]
Re: RegEx question by Anonymous Monk on Dec 26, 2017 at 14:05 UTC
What comes to mind is a few separate `if`-statements. First, gnore the line if it does not contain `nova scotia`. (Least probable so listed first.) Then, perhaps you can simply ignore the line if it does contain `bank` anywhere. Alternatively, you can `split()` the line on `/nova scotia/` and verify that none of the pieces contains `bank`. I have come to dislike fancy-regexes because soon enough some new requirement comes along that messes them up, leaving the poor programmer to make a now-complicated change, and to somehow verify, not only that it handles the new requirement, but also that it continues to handle the old one(s).	[reply]
Re^2: RegEx question by Your Mother (Archbishop) on Dec 26, 2017 at 16:26 UTC
Five days after good real answers, this code-free, poor advice written without reading the thread arrives. One should never make the least likely operation the one always run. The regex is meant to be used in two languages so a coded logic tree isn't desired. Options exist in the construction of code between 2x4s and the JPL.	[reply]