From: Robert Klemme <bob.news@...>
Date: 2005-01-10T17:21:21+09:00
Subject: Re: brute force string search


"Martin Pirker" <crf@sbox.tu-graz.ac.at> schrieb im Newsbeitrag
news:41e1d123$0$11610$3b214f66@aconews.univie.ac.at...
> Hi...
>
> given: String of several Mb
> problem: find the lines in String containing "xyz"
>
>
> Idea 1:
> String.scan(/.*xyz.*/) -> ~10s runtime

Did you try the anchored version?  Did it make a difference?

>> s = "a\nbxyz\ncxyz\nd"
=> "a\nbxyz\ncxyz\nd"
>> s.scan(/^.*xyz.*$/)
=> ["bxyz", "cxyz"]

Did you try the block form, i.e.,

s.scan(/^.*xyz.*$/) {|m| # work with m}

> Idea 2:
> String.grep(/.*xyz.*/) -> ~3s (but gives the \n too)
>
> Idea 3:
> loop String.index("xyz",lastmatch+3)
> loop results array and grep in match area for line
> -> 0,5s
>
>
> C extension the only faster option left? :-)

Im not sure whether you will gain much, because certain things will have
to be done anyway: allocation of the result array, of the match strings
and expansion of the result array.

Kind regards

    robert