Regex question

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

Hello all, Can someone tell me why the following regex does not work? What I want is just the date information "1st day of November 2005". Thanks a bunch!


my $aa = "Accepted and agreed to by Jerry Smith as of the 1st day of N
+ovember 2005.";

if (     $aa =~ m{
    (?:as of)*
    (?:\s|\n)*    #for any possible space or line end    
    (?:this|the)*
    (?:\s|\n)*    #for any possible space or line end            
    (
    (?:[1-31])      #day, 1 to 31
    (?:st|nd|rd|th) #for all the cases like 1st, 2nd, 3rd, 25th
    (?:\s|\n)*      #for any possible space or line end
    (?:day of)*
    (?:\s|\n)*      #for any possible space or line end
    (?:January|February|March|April|May|June|July|
    August|September|October|November|December)
    (?:\s|\n)*    #for any possible space or line end
    (?:\,|\s)*      #for any possible space or comma end
    (?:\s|\n)*      #for any possible space or line end
    ([1,2]\d{2,3})  #year
    )
    (?:\,|\.|\s)*
    (?:[ ]{2}|\n|\z)*   #Two spaces, newline, or string end
    }ixmsg)    {
                
        print $1;
    }
[download]

It works now. I am really grateful to davorg, Fletch, johngg, aggianni,anno, and shmem for all your helpful comments and suggestions. I just get to know PERL for about two weeks and it is great that I find this forum and so many nice and helpful experts here.

Comment on Regex question Download Code

Replies are listed 'Best First'.
Re: Regex question by davorg (Chancellor) on Mar 22, 2007 at 14:36 UTC
`(?:[1-31]) #day, 1 to 31` [download] That's not doing what you think it is. What you have there is functionally equivalent to `[123]` (or `[1-3]`). A simple replacement would be `\d+` (or, perhaps, `\d\d?`). But that also matches many invalid dates. Update: Fletch fills in details that I was too lazy to :-) -- See the Copyright notice on my home node. "The first rule of Perl club is you do not talk about Perl club." -- Chip Salzenberg	[reply] [d/l] [select]
Re^2: Regex question by Fletch (Bishop) on Mar 22, 2007 at 14:47 UTC
`(?:[1-9]\|1[0-9]\|2[0-9]\|3[01])` is a less simple replacement that would accept fewer invalid dates (e.g. you'd still match against `31st day of February`; then again you'd want to do that validation further up the pipe anyhow).	[reply] [d/l]
Re^3: Regex question by johngg (Canon) on Mar 22, 2007 at 15:17 UTC
Another, lazier, way is to use an array and localize the default list separator `@days = ( 1 .. 31 ); local $" = q{\|}; m{(?:@days)};` [download] You'd probably want to limit the `local $"` inside a code block to avoid unexpected side effects elsewhere. Cheers, JohnGG	[reply] [d/l] [select]
Re: Regex question by agianni (Hermit) on Mar 22, 2007 at 14:44 UTC
Be careful when you use the x modifier on the match. It makes your pattern easier to read, but you need to explicitly include all of your white space characters. So `(?:as of)` needs to be `(?:as\sof)`, for example. `perl -e 'split//,q{john hurl, pest caretaker}and(map{print @_[$_]}(joi +n(q{},map{sprintf(qq{%010u},$_)}(2*23074993,51016415261,75979 +36997,13177145131,3*26789167*181))=~/\d{2}/g));'` [download]	[reply] [d/l] [select]
Re: Regex question by Anno (Deacon) on Mar 22, 2007 at 14:58 UTC
Apart from the comments that you have already got, you seem to use the "``" quantifier rather indiscriminately. Are you sure you know what it does in a regex? What you specify as "`(?:\s\|\n)`" is equivalent to a single "`\s`". "`\n`" is already included in the set of characters matched by "`\s`". Anno Update:* Minor corrections	[reply] [d/l] [select]
Re: Regex question by shmem (Chancellor) on Mar 22, 2007 at 15:12 UTC
What I want is just the date information "1st day of November 2005" No, you don't, you also want the date string to be prepended by "as of" etc but then you don't care. (?:shrug)*. If you just wanted the date string, why not just: `$_ = "Accepted and agreed to by Jerry Smith as of the 1st day of November 2005."; print $1 if /(\d+(?:st\|nd\|rd\|th)\s+day\s+of\s+[ADFJMNOSabceghilmnoprst +uvy]+\s+\d+)/ms; __END__ 1st day of November 2005` [download] which would also match "35th day of ANNO 0007", but the matched date strings need to be validated anyways... --shmem _($_=" "x(1<<5)."?\n".q·/)Oo. G°\ / /\_¯/(q / ---------------------------- \__(m.====·.(_("always off the crowd"))."· ");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}	[reply] [d/l]