Ruby, remove last N characters from a string?

If the characters you want to remove are always the same characters, then consider chomp:

'abc123'.chomp('123')    # => "abc"

The advantages of chomp are: no counting, and the code more clearly communicates what it is doing.

With no arguments, chomp removes the DOS or Unix line ending, if either is present:

"abc\n".chomp      # => "abc"
"abc\r\n".chomp    # => "abc"

---

From the comments, there was a question of the speed of using #chomp versus using a range. Here is a benchmark comparing the two:

require 'benchmark'

S = 'asdfghjkl'
SL = S.length
T = 10_000
A = 1_000.times.map { |n| "#{n}#{S}" }

GC.disable

Benchmark.bmbm do |x|
  x.report('chomp') { T.times { A.each { |s| s.chomp(S) } } }
  x.report('range') { T.times { A.each { |s| s[0...-SL] } } }
end

Benchmark Results (using CRuby 2.13p242):

Rehearsal -----------------------------------------
chomp   1.540000   0.040000   1.580000 (  1.587908)
range   1.810000   0.200000   2.010000 (  2.011846)
-------------------------------- total: 3.590000sec

            user     system      total        real
chomp   1.550000   0.070000   1.620000 (  1.610362)
range   1.970000   0.170000   2.140000 (  2.146682)

So chomp is faster than using a range, by ~22%.

Ruby 2.5+

As of Ruby 2.5 you can use delete_suffix or delete_suffix! to achieve this in a fast and readable manner.

The docs on the methods are here.

If you know what the suffix is, this is idiomatic (and I'd argue, even more readable than other answers here):

'abc123'.delete_suffix('123')     # => "abc"
'abc123'.delete_suffix!('123')    # => "abc"

It's even significantly faster (almost 40% with the bang method) than the top answer. Here's the result of the same benchmark:

                     user     system      total        real
chomp            0.949823   0.001025   0.950848 (  0.951941)
range            1.874237   0.001472   1.875709 (  1.876820)
delete_suffix    0.721699   0.000945   0.722644 (  0.723410)
delete_suffix!   0.650042   0.000714   0.650756 (  0.651332)

I hope this is useful - note the method doesn't currently accept a regex so if you don't know the suffix it's not viable for the time being. However, as the accepted answer (update: at the time of writing) dictates the same, I thought this might be useful to some people.

irb> 'now is the time'[0...-4]
=> "now is the "