From: "David A. Black" <dblack@...>
Date: 2009-02-08T05:53:56+09:00
Subject: Re: cannot remove multiple space> nuts!

Hi --

On Sun, 8 Feb 2009, Tom Cloyd wrote:

> David A. Black wrote:
>> Hi --
>> 
>> On Sat, 7 Feb 2009, Tom Cloyd wrote:
>> 
>>> Tom Cloyd wrote:
>>>> I'm baffled by this strange outcome - I cannot reduce multiple spaces 
>>>> from a text file. This isn't just a regex problem, somehow. I'm failing 
>>>> to grasp something essential, but don't know what it is. All help 
>>>> appreciated, as usual!
>>>> 
>>>> Here is a demo of my problem, in which I try two different ways, and both 
>>>> fail:
>>>> 
>>>> === code ===
>>>> # h2t.rb
>>>> 
>>>> def main
>>>>  # conversion table spec
>>>>  conv = [
>>>>  [ '<h1>', 'h1. ' ], [ '<h2>', 'h2. ' ], [ '<h3>', 'h3. ' ],
>>>>  [ '<h4>', 'h4. ' ], [ '<h5>', 'h5. ' ], [ '<h6>', 'h6. ' ],  [ 
>>>> /<\/h\d>/, '' ],
>>>>  [ " +", ' ' ]]  # <= this last array element should do the trick, but 
>>>> doesn't
>>> Ouch. THIS - [ / +/, ' ' ], substituted for [ " +", ' ' ] above fixes it. 
>>> I'm going blind, obviously.
>> 
>> Just for fun, here's another way to write the method:
>> 
>> def main
>>  data = File.read("tom.txt")
>>  data.gsub!(/<(h[1-6])>/, "\\1. ")
>>  data.gsub!(/<\/h\d>/, "")
>>  data.squeeze!(' ')
>>
>>  open("tom.out", "w") {|f| f.write(data) }
>> 
>> end
>> 
>> I think that does the same thing. Tweak to taste :-)
>> 
>> 
>> David
>> 
> That's beautifully economical, and reveals a far better grasp of regex than I 
> was able to attain last night. However, I'm having trouble with this line:
>
> data.gsub!(/<(h[1-6])>/, "\\1. ")
>
> It certain works, but I don't grasp the "\\1. " part. I haven't yet found 
> anything that might shed light on this magic. How does it retain the 'h' and 
> whatever digit follows it? It looks somehow like "\\" == retain matched 
> alpha, and the "1" does the same for matched digits, but I really haven't a 
> clue. Can you elucidate just a bit?

The \\1, \\2, etc. in the replacement string are pegged to the
parenthetical captures. "\\1. " means: the first capture (which is h
plus a digit), a period, and a space.

They work in single-quoted strings too, but there they're just \1, \2,
etc. There's some explanation in the ri docs for String#gsub.


David

-- 
David A. Black / Ruby Power and Light, LLC
Ruby/Rails consulting & training: http://www.rubypal.com
Coming in 2009: The Well-Grounded Rubyist (http://manning.com/black2)

http://www.wishsight.com => Independent, social wishlist management!