From: Robert Klemme <shortcutter@...>
Date: 2010-08-30T22:45:35+09:00
Subject: Re: Speed issues iterating over chars

2010/8/30 Martin Hansen <mail@maasha.dk>:
>> You could speed up calculation of the regexp by placing this in the
>> default block of a Hash, e.g.
>>
>> RXS = Hash.new {|h,cutoff| h[cutoff] =
>> /(?:#{(0..(BASE_SOLEXA+cutoff)).map{|ch|"\\u00%02x" %
>> ch}.join("|")})+/
>>
>> and then
>>
>> seq.gsub! RXS[cutoff] do |m|
>>   m.downcase!
>> end
>
>
> This could be a good idea, but I am not entirely sure if it will work -
> at least the current suggestion gives an empty RXS hash. And I am unsure
> about what this do? map{|ch|"\\u00%02x" % ch}

Just try it out in IRB.  This creates a regexp with unicode names.
You could however use Fixnum#chr instead, e.g.

#{ (0..(BASE_SOLEXA + cutoff)).map {|ch| Regexp.escape(ch.chr)}.join("|") }

This gives shorter and more readable strings.

> Thinking about this problem I wonder if it would be an idea to create a
> mask with transliterate and then use a bitwise operation to downcase.
> IIRC you change the case with a bitwise | operation with ' '. It doesn't
> look like Ruby have any methods bitwise operations on strings.

AFAIK there are none.  But you can do

irb(main):011:0> s = "aaa"
=> "aaa"
irb(main):012:0> s[0] = (s[0].ord | 0x0F).chr
=> "o"
irb(main):013:0> s
=> "oaa"

Kind regards

robert

-- 
remember.guy do |as, often| as.you_can - without end
http://blog.rubybestpractices.com/