RegExp Matched Values

[rpike]

Hi All, I have a regular expression in place that does a check on a string coming in from a form. The regular expression is as such : $string =~ m/^(frm)(\w+)(txt)?$/; What I was hoping to accomplish was to have $1, $2, and $3 populated with the values in the brackets. When I run this and the string is something like frmNametxt the $1 gets set to frm, $2 gets set to Nametxt and $3 to nothing. What (if anything) can I do to force the code to evaluate and store txt into $3? If txt is there I want it put into $3. Any help would be most appreciated. Thanks. Rob

[jethro]

Use $string =~ m/^(frm)(\w+?)(txt)?$/;. This works because \w+? tries to match the smallest possible string now

Note this works only because the string is anchored at the end. Otherwise \w+? would always match only one character and (txt)? the empty string.

[AnomalousMonk]

$string =~ m/^(frm)(\w+?)(txt)?$/;

There's a problem here in that if  'txt' may not be present (as  (txt)? implies) and an overall match is still desired,  $3 will be undefined because the third capture group need not match at all. A different approach will yield an empty string instead.

>perl -wMstrict -le
"for (@ARGV) {
   m/^(frm)(\w+?)(txt)?$/;
   print qq{'$_' ($1) ($2) ($3)};
   }
 print '-----------';
 for (@ARGV) {
   m/^(frm)(\w+?)(txt|)$/;
   print qq{'$_' ($1) ($2) ($3)};
   }
"  frmNametxt  frmNameFoo
'frmNametxt' (frm) (Name) (txt)
Use of uninitialized value $3 in concatenation (.) or string at ...
'frmNameFoo' (frm) (NameFoo) ()
-----------
'frmNametxt' (frm) (Name) (txt)
'frmNameFoo' (frm) (NameFoo) ()

[AR]

In this very specific example, you can change the \w+ into the lazy version: \w+?. If you're looking for a solution to a more general problem, this may not be appropriate.