From: 7stud -- <bbxx789_05ss@...>
Date: 2009-08-05T17:35:38+09:00
Subject: Re: remove commas from string

Jason Lillywhite wrote:
> I have following string:
> 
> s = "B747-400, 8,357 miles, 561 mph, 4 Pratt & Whitney PW 4056
> turbofans, 56,000 lbs."
> 
> I want to remove the comma only from the numbers (8,357 miles and 56,000
> lbs) separating the thousands. I want the string to read as follows:
> 
> "B747-400, 8357 miles, 561 mph, 4 Pratt & Whitney PW 4056 turbofans,
> 56000 lbs."
> 
> I thought this following regex would do the trick because it
> successfully isolated the right commas in rubular.com
> 
> s.gsub(/\d+(,)\d+/, "")
> 
> It turns out that my regex removes the entire number, not just the
> comma.
> 
> Am I wrong in saying that my regex searches for 1 or more numbers
> surrounding a comma and replaces just the comma with ""?
> 

Yes, you are wrong:

s = "|yes|"
puts s.gsub(/y(e)s/, "")

--output:--
||

gsub() replaces the whole match with the specified replacement.  gsub() 
does not pick out a parenthesized group in the regex and replace that 
with the specified replacement .  However, there is a block form of 
gsub:

s = "yes, 1,234, yes, 4,567"

result = s.gsub(/(\d),(\d)/) do |match|
  "#{$1}#{$2}"
end

puts result

--output:--
yes, 1234, yes, 4567

Inside the block, $1, $2, $3, etc. refer to the matches for each 
parenthesized group in the regex.  The return value of the block is used 
as the replacement.
-- 
Posted via http://www.ruby-forum.com/.