CODEWITHSUNDEEP
rubyarraysperformance
Noel Frostpaw
Noel Frostpaw

Reputation: 4009

Ruby : %w performance

perhaps a simple question, but we are discussing about whether it better to use this snipper:

if %w(production staging).include?(Rails.env)

versus

if ["production","staging"].include?(Rails.env)

We just want to understand which is the most performant way, ignoring the sytax suggering from Ruby. From what I could on the web, the %w literal seems to be a shorthand to string.split on the provided whitespace string.

But which one is actually the fastest?

p.s : A source for the answer would be appreciated.

Upvotes: 2

Views: 406

Answers (2)

Michael Kohl
Michael Kohl

Reputation: 66867

Here's what %w and %W do, taken directly from parse.y (with ommissions):

case '%':
[snip]
  switch (c) {
    [snip]
    case 'W':
      lex_strterm = NEW_STRTERM(str_dword, term, paren);
      do {c = nextc();} while (ISSPACE(c));
      pushback(c);
      return tWORDS_BEG;

    case 'w':
      lex_strterm = NEW_STRTERM(str_sword, term, paren);
      do {c = nextc();} while (ISSPACE(c));
      pushback(c);
      return tQWORDS_BEG;

Considering it's implemented on the parser level, I wouldn't worry too much about the performance.

Upvotes: 7

shiroginne
shiroginne

Reputation: 89

I've done some test on my c2d:

ruby -e "10000000.times { ['one', 'two'].include?('two')}"  
8.04s user 0.05s system 90% cpu 8.912 total

ruby -e "10000000.times { %w(one two).include?('two')}"  
8.03s user 0.05s system 93% cpu 8.608 total

Upvotes: 3

Related Questions