Trying to create this type of regular expression...

[T12]

I am trying to create a regular expression that will find matches for select keywords both in front and after different phrases. I'll try to explain it.. for example:

I have three keywords: word1|word2|wor d3 that I want matched with three different phrases phrase1|phrase2 |phrase3

So I'd like the following to "hit"

word1 phrase1
phrase2 word 1
etc..

But I do not want word1,2,3 or phrase1,2,3 to be matched alone, and I'm trying to have this all done in a single string. Hopefully this makes sense to some of you out there... thanks!!

[miller]

Greetings and Welcome,

Why not just keep this simple and use two regular expressions?

[CODE=perl]
while (<DATA>) {
if ($_ =~ /word1|word2|wor d3/ && $_ =~ /phrase1|phrase2 |phrase3/) {
print "Match found: $_";
}
}

__DATA__
word1 no phrace
no phrase word2
word1 phrase1
phrase2 word3
phrase3 phrase2 phrase1
word1 word3 word2
phrase2 phrase3 word1
[/CODE]

Output:

>perl scratch.pl
Match found: word1 phrase1
Match found: phrase2 word3
Match found: phrase2 phrase3 word1

Yes it is possible to do this in a single regular expression. But why complicate things? However, there is one small bug in the above code and that's the fact that if a word is within a phrase, that will always match.

One way to do the equivalent of the above in a single regex using zero-width positive lookahead assertions. You can read about them here:

perldoc perlretut

- Miller

[T12]

Thank you for the information. I've had a look at the site, but unfortunately I don't really understand how to compose the single string I need to get these type of matches. I'm still curious about how to accomplish this in one line. Would you be able to post the complete string? I'll continue to research and hopefully figure it out at some point...

[miller]

Yes, I'll give you the method that I was talking about. However, I discourage you from using this until you actually know how it works. You'll even notice that it actually takes more code to do this in one regex. Nevertheless, here's the example:

[CODE=perl]
while (<DATA>) {
# if ($_ =~ /word1|word2|wor d3/ && $_ =~ /phrase1|phrase2 |phrase3/) {
if ($_ =~ /^(?=.*?(?:word1 |word2|word3))( ?=.*?(?:phrase1 |phrase2|phrase 3))/) {
print "Match found: $_";
}
}

__DATA__
word1 no phrace
no phrase word2
word1 phrase1
phrase2 word3
phrase3 phrase2 phrase1
word1 word3 word2
phrase2 phrase3 word1
[/CODE]

- Miller