regexp to check if file name matches pattern

[smanicka]

Hello monks,I have a file names of the pattern SYSBLAHBLAH%BLAHBLAH%BLAHBLAH%timedate.pdf

Could anyone help me write a regexp to check if a given filename matches this pattern?

I figure i have to say =~/^SYS something /i but i am not sure how exactly to match up with the first 2 percentage signs, the time date stamp is not of concern.Any help would be appreciated.

i need to do : if filename matches pattern, do action .else skip

i am attempting the regex as =~m/^SYS\w%\w%\w.*/i

Would this work? Thanks

- Sandhya

[davido]

Perhaps this:

/^SYS(?:[^%]+%)+[^.]*\.pdf/i

Matches SYS followed by any non % character(s) that are followed by a % (can be repeated), followed by just about anything except a dot (.) (that's where the date field goes), followed by dot pdf.

Update: /i added, because when I see SYS and .pdf (upper, and lower) I start thinking that case probably shouldn't be allowed to foul up a match.

[smanicka]

the systemblahblah can contain a period unfortunately

[davido]

I didn't preclude the dot from "systemblah%blah%", I precluded dot from the datestamp. Did you try it yet? There may be other errors, but that one isn't. ;)

[Marshall]

An alternate solution (there are many):

$file= "SYSB_A31.GROUPINS%RDEBLIST%REAPPLY%2008.12.16.20.31.35.pdf";
if ($file =~ /^SYS[\w.]+%[\w.]+%[\w.]+%[0-9.]+.pdf/)
{print "yes"}
else
{print "no"}
#prints yes

A set of characters is enclosed in []. \w are the "word chars" and we add the "dot" to that set by \w.+. Same idea with 0-9.+. Other solution that has [^%]+ is good too. That means anything except %. Take your pick.

[smanicka]

Hello Dave, it doesn't work.I tried:

#!/usr/bin/perl
 $file= "SYSB_A31.GROUPINS%RDEBLIST%REAPPLY%2008.12.16.20.31.35.pdf";
if ($file=~ /^SYS(?:[^%]+%)+[^.]*\.pdf/i){
print "yes";
} else{
print "no";
}

and it prints out no. Thanks Smanicka

[davido]

Lol, I had no way of knowing that your date field was "dot delimited." My solution would naturally reject that. Here's a new approach that should do the trick given that new, previously unmentioned requirement.

/^SYS(?:[^%]+%){2,}[\d.]+\.pdf/i

Let me know how that works out. This time I'm specifically allowing the 'dot' delimiter in the date field. This solution is going to result in some backtracking though. It might be better written as:

/^SYS(?:[^%]+%){2,}(?:\d+)(?:.\d+)+\.pdf/i

This will get less backtracking, and thus should be a little more efficient, if that matters.

[smanicka]

tried {2,} instead of + as suggested.since prints out no

[Grey Fox]

Howdy;
I have posted in the past a great tool for explaining regexe's called My REGEX Tester. I have a link to it in Found new online Regex tool.
Plus there is a nice website called Regular-Expressions Quickstart. Which teaches all of the basics of REGEX.


Teach a person to fish and you feed them for a lifetime.