newline-strip.rb

#! /usr/bin/ruby

#
#  I've got a string containing newlines. I've also got a regex that I
#  want to match at the start of the string, but NOT at the start of a
#  line.
#
#  Contrived example:
#    - strings: "first string\nshould match", "second string should not match\nfirst string"
#    - regex:   /^first string/
#  We want to match only on the first string, not on the second string.
#
#  Seems like the best way to do this, assuming my regex does not
#  contain newlines, is to remove everything after the first newline
#  in the string.
#
#  But, then, what's the fastest way to remove everything after the first newline?
#

require 'benchmark'
require 'open-uri'

#doc = open("http://www.slashdot.com/").read # large doc
doc = "   first string\nshould match" # small doc

puts "length: #{doc.length}"

N = 100_000
Benchmark.bm(15) do |bm|
  bm.report("gsub:")        { N.times do ; doc.gsub("\n.*","") ; end }
  bm.report("delete:")      { N.times do ; doc.delete("\n") ; end }
  bm.report("slice(re):")   { N.times do ; doc.slice(/^[^\n]+/) ; end }
  bm.report("index+slice:") { N.times do ; j = doc.index("\n") ; doc.slice(0,j) ; end }
end

## where 'doc' is large
# length: 108602
#                      user     system      total        real
# gsub:           31.900000   9.030000  40.930000 ( 40.960776)
# delete:         23.130000   0.080000  23.210000 ( 23.241717)
# slice(re):       0.330000   0.020000   0.350000 (  0.347720)
# index+slice:     0.130000   0.020000   0.150000 (  0.148842)

## where 'doc' is small
# length: 25
#                      user     system      total        real
# gsub:            0.120000   0.020000   0.140000 (  0.132871)
# delete:          0.150000   0.010000   0.160000 (  0.158218)
# slice(re):       0.170000   0.020000   0.190000 (  0.185259)
# index+slice:     0.100000   0.020000   0.120000 (  0.125409)