question when learning P::RD?

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

Hello Monks,

Long time no see! Recently I have to parse a file and export to another format. the original file is like this:

Y 034309201607258   1        #Y means header
Q     02751VACANT / CLOSE    #comment
G   ....            # inner head
...
...
...                          
T RCRC   0810010
T RDRD   0810010
A 22OP
A 13O 
A 12O 
P 3472CHSK00010014 #P means inner trailer
Z 034309201607258  #Z means trailer
[download]

I took this weekend to learn P::RD, and seems a little bit to understand it. but when I tried some snippet to run, it's hard to succeed as I image:

use Parse::RecDescent;
  use IO::All;

  my $text = io("xxx.DAT")->slurp;
  # Create and compile the source file
  $parser1 = Parse::RecDescent->new(q(
    startrule : HeadRule /.+/ms TrailerRule 

    HeadRule : /^Y.+$/m
    TrailerRule: /^Z.+$/m
  ));

  $parser2 = Parse::RecDescent->new(q(
    startrule : HeadRule 

    HeadRule : /^Y.+^Z.+$/ms

  ));


# Test it
  print "Valid data\n" if $parser1->startrule($text); #no!
  print "Valid data\n" if $parser2->startrule($text); #yes
[download]

why parser1 is failed? and in order to learn P::RD, Could anyone tell me if some module parsing files( like xml or html etc> on cpan make use of P:RD?

Thanks

I am trying to improve my English skills, if you see a mistake please feel free to reply or /msg me a correction

Comment on question when learning P::RD? Select or Download Code

Replies are listed 'Best First'.
Re: question when learning P::RD? by duelafn (Parson) on Aug 28, 2016 at 17:26 UTC
If you add `$::RD_TRACE = 1;` to your script, P::RD will print out a trace of the parse which can help with the debugging. In this case, it shows that the `/.+/ms` pattern gobbles up "`Z 034309201607258`" (P::RD can't backtrack into the pattern, it will only backtrack whole rules). Your middle rule should leave the "Z" line at the end. For instance: #!/usr/bin/perl use strict; use warnings; use 5.014; use Parse::RecDescent; # $::RD_TRACE = 1; my $text = <<'CONTENT'; Y 034309201607258 1 Q 02751VACANT / CLOSE G .... ... ... T RCRC 0810010 T RDRD 0810010 A 22OP A 13O A 12O P 3472CHSK00010014 Z 034309201607258 CONTENT my $parser1 = Parse::RecDescent->new(<<'GRAMMAR'); startrule : HeadRule OtherRule(s) TrailerRule HeadRule : /^Y.+$/m TrailerRule: /^Z.+$/m OtherRule: /^(?![YZ]).*$/m GRAMMAR print "Valid data\n" if defined($parser1->startrule($text)); [download] Also, your parser2 shows why you should almost always end your top-level rule with a `/\Z/` pattern so that parsing fails if P::RD doesn't consume the whole string. Good Day, Dean	[reply] [d/l] [select]
Re: question when learning P::RD? by Laurent_R (Canon) on Aug 28, 2016 at 10:03 UTC
Hi xiaoyafeng, I am not sure you really need a parser for such a simple format. Simple regexes on individual lines should probably do the work. I haven't been using the Parse::RecDescent recently, so this may have no consequence, but maybe you could try to remove the spaces before the colons (":" characters) in your rule definitions.	[reply]
Re^2: question when learning P::RD? by xiaoyafeng (Deacon) on Aug 28, 2016 at 12:30 UTC
Laurent_R, Thanks for your reply, the format of original file is not as simple as I showed. it's about 10M size and various types nest in it.(like ^M xxxx, ^D xxxx etc). besides, it's a good chance to learn knowledge of parsing language to me. I want to have a try. ;) I am trying to improve my English skills, if you see a mistake please feel free to reply or /msg me a correction	[reply]
Re: question when learning P::RD? by ikegami (Patriarch) on Aug 29, 2016 at 15:08 UTC
`/.+/ms` matches the rest of the file, so there's nothing for `TrailerRule` to match.	[reply] [d/l] [select]