Deleting everything before a string

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

Replies are listed 'Best First'.
Re: Deleting everything before a string by I0 (Priest) on Dec 21, 2000 at 07:04 UTC
`$content =~ s/.\+_(.?)\+_.*/$1/s; #or $content = (split/\+_/,$content)[1];` [download]	[reply] [d/l]
Re: Re: Deleting everything before a string by maverick (Curate) on Dec 21, 2000 at 09:07 UTC
I would prefer `@content = ($content =~ /\+_(.*?)\+_/sg);` [download] since it would work for more that occurance of the replacement string in the file. /\/\averick	[reply] [d/l]
Re: Re: Re: Deleting everything before a string by I0 (Priest) on Dec 21, 2000 at 19:19 UTC
Although it is unclear to me how Anonymous Monk would want to handle in that case. Perhaps Anonymous Monk would prefer `@content = ($content =~ /\+_(.?)(?=\+_)/)sg);` or `/\+_(.)\+_/s`	[reply] [d/l] [select]
Re: Re: Deleting everything before a string by repson (Chaplain) on Dec 21, 2000 at 07:08 UTC
And then of course writing the new $content over the file you read in from. On the other hand if you are dealing with a large file then I would suggest searching line by line (or chunk by chuck) for +_...+_ and then storing that, closing the file, writing the found information to a new file of the same or different name.	[reply]
Re: Re: Re: Deleting everything before a string by I0 (Priest) on Dec 21, 2000 at 19:28 UTC
Or perhaps setting $/ to '+_' and $^I to '.bak'	[reply]
Re: Deleting everything before a string by fundflow (Chaplain) on Dec 21, 2000 at 08:06 UTC
Instead of adding extra stuff and then erasing it, you could just as well do: `while (<MYFILE> ) { $content .= $_ if /start-string/ .. /end-string/ }` [download]	[reply] [d/l]
Re: Deleting everything before a string by cat2014 (Monk) on Dec 21, 2000 at 09:28 UTC
the easiest way is just `s/^(.*)\+_/\s\+_$1\+_/s` of course, there are a few caveats with that: i don't believe that . matches the new line character, (unless you have the s at the end of the s///) and you will only end up with one space for all the stuff before the + that you replaced. did you want just one white space character, or 1 for every character that you replaced? and i just realized that you want to delete after the +, too. the first answer is much better than my quick regex. would tr/// work for this?	[reply] [d/l]
Re: Deleting everything before a string by Cybercosis (Monk) on Dec 21, 2000 at 15:05 UTC
Okay, this is going to be evil. (note to self: don't program in Perl after learning Standard ML...) but it (might) be somewhat faster, since there's no regex stuff: `my $first_index = index($content, "+_"); my $second_index = index($con +tent, "+_", $first_index); $content = substr($first_index, $first_index - $second_index);` [download] Okay, that wasn't so bad... my Perl instincts almost raged against this, which may not even work: `$content = substr($content, index($content, "+_")+2, index($content, " ++_") - index($content, "+_", index($content, "+_")+1)-4);` [download] shudder damn functional languages... =-)	[reply] [d/l] [select]
Re: Deleting everything before a string by snax (Hermit) on Dec 21, 2000 at 14:59 UTC
From the description of your algorithm, it sounds like you want `$content` to contain only a specific string if that string is contained within the file in question, right? Or do you really mean you want to replace everything else with whitespace, so that `xxxJOHNxxx` would be three spaces, `JOHN`, and then three spaces? In the first case, how about `($file_text =~ /$string_to_match/) && $content = $string_to_match;` [download] There's no need to do a substitution since you already know what you want the value of `$content` to be. Moreover, you probably have a default value in mind if the file fails to contain the match string, in which case you could use the ternary operator to do it: `$content = ($file_text =~ /$string_to_match/) ? $string_to_match : $default_failure_value;` [download] Finally, if you know the file contains the text, and that's what you want in the `$content` variable, just do a simple assignment. Update: In case it's not obvious, I construed the question to imply that the matching string was known beforehand, rather than the delimiters. If the delimiters are what is known beforehand, my advice is, well, pointless.	[reply] [d/l] [select]