regexp match arg1 AND arg2

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

Wise Monks,

In a command line telephonebook search program, people can use any number of arguments which must all match.
Example:

TelSearch.pl bill clinton
[download]

Searches for all entries that have both 'bill' AND 'clinton' in the entrystring (which contains name, alias, room# etc)

Is it possible to make ONE regexp for this purpose? It would be easy if the order of the arguments mattered, but it doesn't. Example:

TelSearch.pl bill clinton
[download]

Needs to find the same entries as

TelSearch.pl clinton bill
[download]

I currently use something like this:

my $flag = 1;
foreach (@ARGV)
{
  if ($entry !~ m/\Q$_\E/)
  {
    $flag = 0;
  }
}
[download]

Comment on regexp match arg1 AND arg2 Select or Download Code

Replies are listed 'Best First'.
Re: regexp match arg1 AND arg2 by broquaint (Abbot) on Feb 17, 2003 at 15:19 UTC
Your predicament is very similar to Dynamic regexp from array values so you'll find many an appropriate answer there. HTH `_________ broquaint`	[reply]
Re: Re: regexp match arg1 AND arg2 by Jaap (Curate) on Feb 17, 2003 at 16:32 UTC
Thank you. Although not one regexp, the other topic had a shorter way to do it: `doSomething() if grep $data =~ /\Q$_/, @vals;` [download] That should match all in no particular order.	[reply] [d/l]
Re: regexp match arg1 AND arg2 by Limbic~Region (Chancellor) on Feb 17, 2003 at 15:33 UTC
There is definately MTOWTDT. Since you have only requested assistance for two arguments, you might try something like: `if ($entry =~ /\Q$ARGV[0]\E/ && $entry =~ /\Q$ARGV[1]\E/) { # Do something }` [download] The problem is that doesn't technically meet your requirements of ONE regexp. This would be another way to do it with ONE regexp: `if ($entry =~ /(\Q$ARGV[0]\E.\Q$ARGV[1]\E\|\Q$ARGV[1]\E.\Q$ARGV[0]\E) +/ { # Do something }` [download] But that is going to be much slower and probably has about a dozen ways in which it won't work My recommendation would be to standardize your data fields (i.e. first name, last name, etc) and then to use command line options to match them. TelSearch.pl -f bill -l clinton or TelSearch.pl -l clinton -f bill Hope this helps! Cheers - L~R Update: Corrected regexp that Enlil pointed out would only match one of the two arguments.	[reply] [d/l] [select]
Re: Re: regexp match arg1 AND arg2 by Jaap (Curate) on Feb 17, 2003 at 16:29 UTC
No i am sorry but people can use any number of arguments that must match. So your solution will not work :(	[reply]
Re: Re: Re: regexp match arg1 AND arg2 by Limbic~Region (Chancellor) on Feb 17, 2003 at 16:35 UTC
Then I suggest you try one of the suggestions that broquaint already pointed out, or you take my advice and structure the data and provide command line options. Cheers - Limbic~Region	[reply]
Re: regexp match arg1 AND arg2 by hv (Prior) on Feb 17, 2003 at 17:36 UTC
If you particularly want a single regular expression, you can get it by using lookahead: `my $re = join ' ', map "(?= .? \Q$_\E)", @ARGV; if ($entry =~ /^ $re/sx) { ... }` [download] That would turn your example arguments into something like: `m{ ^ (?= .? bill ) (?= .*? clinton ) }xs` [download] so it matches if: `from the start fail unless we can find some characters followed by "bill" _and_ (still from the start) fail unless we can find some characters followed by "clinton" succeed` [download] I'm not sure how efficient that is, but it should be ok; if speed is important it would be worth benchmarking against some alternatives. Don't forget that when looking for an exact string match, `index()` can be faster than `m/\Q$str\E/`, and that if you're going to match one string against many regexps `study($string)` can help. Hugo	[reply] [d/l] [select]
Re: Re: regexp match arg1 AND arg2 by Jaap (Curate) on Feb 18, 2003 at 11:27 UTC
Wow this is exactly what i meant. Thank you! It works like a charm.	[reply]
Re: regexp match arg1 AND arg2 by zengargoyle (Deacon) on Feb 17, 2003 at 17:58 UTC
something like this perhaps... `@matches = map {qr/\Q$_\E/} @ARGV; while ($rec = <DATA>) { print $rec if @matches # we have something to look for and ( grep {$rec =~ $_} @matches # the number of matches found ) == @matches; # equals the number specified } __END__ Bill Gates Bill Bonko John Random Quincy Gates` [download] $ perl test.pl Bill Bill Gates Bill Bonko $ perl test.pl Bill Bonko Bill Bonko $ perl test.pl Bonko Bill Bonko perl test.pl Gates Bill Gates Quincy Gates $ perl test.pl Gates Bill Bill Gates	[reply] [d/l]