Re: Re: most efficient regex to delete duplicate words

I'm still in the process of learning regexes, but I would think that

$strint=~s/(\w+)(?:\W\1)+/\1/g;

would be a little faster, as you don't need to store the second parenthesis match thingy, and you don't need to match zero or more copies of the first match, you need to match one or more.

Can someone please correct me or confirm this thought? Thanks.

Michael

PS: I've also been thinking about the /o modifier. Would this be a good place to use it?

Comment on Re: Re: most efficient regex to delete duplicate words Download Code

Replies are listed 'Best First'.
Re: Re: Re: most efficient regex to delete duplicate words by blakem (Monsignor) on Aug 14, 2001 at 11:00 UTC
PS: I've also been thinking about the /o modifier. Would this be a good place to use it? $strint=~s/(\w+)(?:\W\1)+/\1/g; Nope. And heres why. The /o modifier relates specifically to interpolated variables inside of the regex. For instance, the regex: `m/$matchvar/` Is a candidate for the /o modifier, since it will be built from the value of $matchvar. With the /o modifier, the resulting "regex engine" will be cached for future use. The next time we run across this regex, we'll use the same engine, even if $matchvar has changed. The common analogy is that the /o modifier is like a promise. You promise not to change any of the variables in the regex, in return for better performance. If you break your promise, your program will break as well. Your original regex has no interpolated variables, so the regex is only compiled once and we use that copy everytime regardless of the /o modifier. In other words, the /o modifier would be redundant on your regex. -Blake	[reply] [d/l]

Replies are listed 'Best First'.

Re: Re: Re: most efficient regex to delete duplicate words
by blakem (Monsignor) on Aug 14, 2001 at 11:00 UTC

PS: I've also been thinking about the /o modifier. Would this be a good place to use it?
$strint=~s/(\w+)(?:\W\1)+/\1/g;

Nope. And heres why.

The /o modifier relates specifically to interpolated variables inside of the regex. For instance, the regex:

m/$matchvar/

Is a candidate for the /o modifier, since it will be built from the value of $matchvar. With the /o modifier, the resulting "regex engine" will be cached for future use. The next time we run across this regex, we'll use the same engine, even if $matchvar has changed.

The common analogy is that the /o modifier is like a promise. You promise not to change any of the variables in the regex, in return for better performance. If you break your promise, your program will break as well.

Your original regex has no interpolated variables, so the regex is only compiled once and we use that copy everytime regardless of the /o modifier.

In other words, the /o modifier would be redundant on your regex.

-Blake

[reply]
[d/l]