Formatting URL's from text file - not easy for me! 2

[ednit]

I am attempting to reformat the structure of URL's inside a text file, but I am about pulling my hair out.

The urls are like this inside a text file:

[URL unfurl="true"]http://www.domain.com/http://www.somesite.com/index.phphttp://www.[/URL] . .

In other words, every URL is butted up next to the other without a space in between. I want to return the text into an easy to use format like so:

[URL unfurl="true"]http://domain.com[/URL]
[URL unfurl="true"]http://www.example.com[/URL]
[URL unfurl="true"]http://site.com/folder/[/URL]
etc. . .

The code I'm using is this:

open(DATA, "data.txt") or die "you stupid script";
while (<DATA>) {
@links = split(/^http+$/, $_);
print "$links[0]\n";
}
close DATA;

$links[0] prints out the whole @links - not only the first link.

I'm not a programmer really, more of a hack that just knows a few things. . . but this has had me stopped up for over 2 hours. Any help would be appreciated.

Thanks.

 

[MillerH]

Your problem is in the regular expression. Try this one instead:

@links = split [COLOR=blue]'(?=[URL unfurl="true"]http://)'[/URL][/color], $_;

 

[ednit]

Thank you so much! That worked perfectly, and I've never seen something formatted like that before. . .I might never have figured it out.

The finished code, in case someone else needs to use it:

open(DATA, "data.txt") or die "you stupid script";
while (<DATA>) {
@links = split '(?=[URL unfurl="true"]http://)',[/URL] $_;
foreach $links (@links) { 
print "$links<br>\n"; 
}
}
close DATA;

So maybe I can understand it & learn from it, what does the equal sign & the question mark mean in this instance? Basically what does '?=matchtext' mean?

This part is not imperitive to the code, only I don't understand the why or what of the above regular expression that fixed my issue.

An explination is not necessary, but a big thank you again.

 

[MillerH]

(?=regexp) is a zero width positive lookahead assertion. You can read about it and other types of assertions in perlretut:

http://perldoc.perl.org/perlretut.html#Looking-ahead-and-looking-behind

 

[ednit]

Thank you again, and especially for that link. I've never heard or read a reference of 'zero width positive lookahead assertion' before that I know of.


-Rob

 

[PC888]

Hello MillerH,

Your code works perfectly with the look-ahead extension.

While I was trying to add the modifier 'i' to make it case insensitive, I succeeded in doing it if I use '/' as a delimiter, but it leads to awkward code.

I understand that a prefix operator is required if we do not use / as the delimiter. Would you suggest to me what should be done using the single quote as a delimeter and the i modifier?

This is what I've got so far:

$line="

http://www.abc.comHTTP://www.def.comhttp://www.ghi.com";

#@array=split '(?=

http://)',

$line; # case sensitive
@array=split /(?=http:\/\/)/i, $line; # case insensitive
#@array=split '(?=

http://)'i,

$line; # compilation error
foreach $i (@array){printf("%s\n",$i);}

 

[PC888]

The link you provided gave the answer, I only need the letter m before delimeters other than /. Thus the line should now read, and it works perfectly.

$line="

http://www.abc.comHTTP://www.def.comhttp://www.ghi.com";

@array=split m'(?=

http://)'i,

$line;
foreach $i (@array){printf("%s\n",$i);}

 

[MillerH]

I love it when people answer their own questions. It's just soo .... efficient.

I believe that you came up with the best solution for your case-insensitivity problem. However, just to introduce an alternate method, it also would have been possible to embed the modifier in the regex.

my $line = "[URL unfurl="true"]http://www.abc.comHTTP://www.def.comhttp://www.ghi.com";[/URL]
my @array = split '[COLOR=green](?i)[/color](?=[URL unfurl="true"]http://)',[/URL] $line;
print "$_\n" foreach @array;

This can be read more about in perldoc as well:

http://perldoc.perl.org/perlretut.html#Embedding-comments-and-modifiers-in-a-regular-expression

 

[PC888]

While answering one's own question sounds a little... studpid, for lack of a better word, it is sure efficient, as you said, saving everyone else from having to work on the problem if an answer has already been found.

I vote your answer with the embedded modifier as the best, as it can be varied, turned on and off through the expression. Thanks again.