Puzzling word count after removing tags

[Ramnarayan]

Hi folks,

I am trying to find the # of words in a xml file. The tricky thing here is we need to count the words which are not in the tags, but outside the tags. An example here is given:

<editor>Edited by
<persName>
<foreName>John</foreName>
<surname>Barry</surname>
</persName> and
<persName>
<foreName full="init" type="first">E.</foreName>
<foreName type="middle">Gene</foreName>
</persName>
</editor>

Now, I am trying to count the words "Edited by", "John", "Barry", "and", "E." and "Gene" from the above list. As you can see anything starting with "<" and ending with ">" can be called as a tag like <editor>...</editor>.

Here is the script I wrote. But it is counting only one word i.e only "Edited by". Infact it should count two words here. Where am I wrong!

$line =~ s,<[/]?.*?>,,g;
if ($line =~ m/(\s\b(.+)\b\s)/g)
{
++$wc if ( defined $1);
}
else
{
fatal ("cannot pull word from $.:\n $!");
}
print "Word count: $wc\n";

 

[PaulTEG]

Why write your own XML parser, there are a few available on

http://search.cpan.org

HTH
--Paul

It's important in life to always strike a happy medium, so if you see someone with a crystal ball, and a smile on their face ...

 

[figmatalan]

It's not pretty but this works:

$str = "
<editor>Edited by
<persName>
<foreName>John</foreName>
<surname>Barry</surname>
</persName> and
<persName>
<foreName full=\"init\" type=\"first\">E.</foreName>
<foreName type=\"middle\">Gene</foreName>
</persName>
</editor>
";

print "$str\n\n";

$str =~ s/<.*?>//g;
$str =~ s/<\/.*?>//g;
$str =~ s/^\s+//;

@words = split (/\s+/, $str);

print "words = ".@words."\n";

$count = 0;
foreach (@words) {
$count++;
print "$count - $_\n";
}

 

[bardley]

Or, if you want to do it as it comes in from the file,

$wc = 0;
while (<XMLFILE>) {
    chomp;
    s!\s*</?.+?/?>\s*!!g;
    if (/\w/) {  # quick sanity check for any non-whitespace character
        @words = split /\s+/;
        $wc += scalar @words;
    }
}

Also, I could be wrong, but I'm pretty sure strict xml doesn't allow things like

<tag1>someval<tag2></tag2></tag1>

so that may make using an xml parser difficult.

Brad Gunsalus
bardley90@hotmail.com

 

[laserBeam]

When you use the /g , the expression returns a list not a scalar value anymore since you can have multiple matches.

  {
    $line =~ s,<[/]?.*?>,,g;
    @a = ($line =~ m/\w+/g);  
    $wc += scalar(@a); # or $wc += @a;
  }
  print "Word count: $wc\n";

or if you want to be very fancy ;-), you can use an expansion ...

  {
    $line =~ s,<[/]?.*?>,,g;
    $wc += @{[($line =~ /\w+/g)]};
  }
  print "Word count: $wc\n";

Cheers!

 

[duncdude]

why not do the opposite of finding the words you want... by rejecting the ones you don't want

[b]#!/usr/bin/perl[/b]

undef $/;

$xml = <DATA>;

$xml =~ s|<[^>]+>||g;

print $xml;

[blue]__DATA__
<editor>Edited by
  <persName>
    <foreName>John</foreName>
    <surname>Barry</surname>
  </persName> and
  <persName>
    <foreName full="init" type="first">E.</foreName>
    <foreName type="middle">Gene</foreName>
  </persName>
</editor>[/blue]

Kind Regards
Duncan

 

[Ramnarayan]

Thanks for everyone to give their responses. However Duncan's response is very good.
$xml =~ s|<[^>]+>||g;
But you need to pass the output of the $xml to a array and count the words. This worked correctly compared to the pattern search given by others ($xml =~ s|<[^>]+>||g

As for Paul, Bardley is right. XML:arser is hard when you need to calculate the # of words and other indepth analysis of tagging. XML:ARSER really works if when you need to work on the parsing of XML files with certain limitations.

 

[duncdude]

Thanks dude!


Kind Regards
Duncan