I have a big text:
"Big piece of text. This sentence includes 'regexp' word. And this
sentence doesn't include that word"
I need to find substring that starts by 'this' and ends by 'word' but doesn't include word 'regexp'.
In this case the string: "this sentence doesn't include that word" is exactly what I want to receive.
How can I do this via Regular Expressions?
With an ignore case option, the following should work:
\bthis\b(?:(?!\bregexp\b).)*?\bword\b
Example: http://www.rubular.com/r/g6tYcOy8IT
Explanation:
\bthis\b # match the word 'this', \b is for word boundaries
(?: # start group, repeated zero or more times, as few as possible
(?!\bregexp\b) # fail if 'regexp' can be matched (negative lookahead)
. # match any single character
)*? # end group
\bword\b # match 'word'
The \b surrounding each word makes sure that you aren't matching on substrings, like matching the 'this' in 'thistle', or the 'word' in 'wordy'.
This works by checking at each character between your start word and your end word to make sure that the excluded word doesn't occur.
Use lookahead asseterions.
When you want to check if a string does not contain another substring, you can write:
/^(?!.*substring)/
You must check also the beginning and the end of line for this and word:
/^this(?!.*substring).*word$/
Another problem here is that you don't want to find strings, you want to find sentences (if I understand your task right).
So the solution looks like this:
perl -e '
local $/;
$_=<>;
while($_ =~ /(.*?[.])/g) {
$s=$1;
print $s if $s =~ /^this(?!.*substring).*word[.]$/
};'
Example of usage:
$ cat 1.pl
local $/;
$_=<>;
while($_ =~ /(.*?[.])/g) {
$s=$1;
print $s if $s =~ /^\s*this(?!.*regexp).*word[.]/i;
};
$ cat 1.txt
This sentence has the "regexp" word. This sentence doesn't have the word. This sentence does have the "regexp" word again.
$ cat 1.txt | perl 1.pl
This sentence doesn't have the word.
(?! will prevent it from matching any case where the substring appears after this (from the example), even if the substring appears after word. Using a nested (?:(?! per F.J's answer fixes this. I don't have Perl available to test that snippet, but I'm not learning anything by looking at it, and from what I can tell it's not a RegEx solution anyway, since it appears to be manually walking through the string with that while loop. The question's tagged regex not perl.This sentence has the "regexp" word. This sentence doesn't have the word. This sentence does have the "regexp" word again. your code will find no results.