What's the opposite of chr() in Ruby?

Question

In many languages there's a pair of functions, chr() and ord(), which convert between numbers and character values. In some languages, ord() is called asc().

Ruby has Integer#chr, which works great:

>> 65.chr
A

Fair enough. But how do you go the other way?

"A".each_byte do |byte|
   puts byte
end

prints:

and that's pretty close to what I want. But I'd really rather avoid a loop -- I'm looking for something short enough to be readable when declaring a const.

score 84 · Accepted Answer · answered Dec 10 '13 at 17:50

84

If String#ord didn't exist in 1.9, it does in 2.0:

"A".ord #=> 65

answered Dec 10 '13 at 17:50

Rob Cameron

9,674
7
39
42

score 33 · Answer 2 · answered Nov 21 '08 at 14:19

33

In Ruby up to and including the 1.8 series, the following will both produce 65 (for ASCII):

puts ?A
'A'[0]

The behavior has changed in Ruby 1.9, both of the above will produce "A" instead. The correct way to do this in Ruby 1.9 is:

'A'[0].ord

Unfortunately, the ord method doesn't exist in Ruby 1.8.

answered Nov 21 '08 at 14:19

Robert Gamble

106,424
25
145
137

It's unfortunate that the "correct" way in Ruby 1.9 is so long, but at least it'll show up easier in searches for "ord". Thanks for your very detailed answer. – RJHunter Nov 22 '08 at 11:30

score 13 · Answer 3 · answered Nov 21 '08 at 13:44

13

Try:

'A'.unpack('c')

answered Nov 21 '08 at 13:44

dylanfm

6,292
5
28
29

1

Now that Ruby 1.9 has changed the meaning of 'A'[0], this is the more portable method. – AShelly Nov 22 '08 at 01:11

Martin Dorey · Answer 4 · 2011-05-27T21:14:58.150

I'd like to +1 dylanfm and AShelly's comment but add the [0]:

'A'.unpack('C')[0]

The unpack call returns an Array containing a single integer, which is not always accepted where an integer is wanted:

$ ruby -e 'printf("0x%02X\n", "A".unpack("C"))'
-e:1:in `printf': can't convert Array into Integer (TypeError)
    from -e:1
$ ruby -e 'printf("0x%02X\n", "A".unpack("C")[0])'
0x41
$

I'm trying to write code that works on Ruby 1.8.1, 1.8.7 and 1.9.2.

Edited to pass C to unpack in uppercase, because unpack("c") gives me -1 where ord() gives me 255 (despite running on a platform where C's char is signed).

score 5 · Answer 5 · 2014-04-05T06:12:33.730

5

Just came across this while putting together a pure Ruby version of Stringprep via RFCs.

Beware that chr fails outside [0,255], instead use 1.9.x - 2.1.x portable replacements:

[22] pry(main)> "\u0221".ord.chr
RangeError: 545 out of char range
from (pry):2:in 'chr'
[23] pry(main)> x = "\u0221".unpack('U')[0]
=> 545
[24] pry(main)> [x].pack('U')
=> "ȡ"
[25] pry(main)>

edited Apr 05 '14 at 06:12

answered Apr 05 '14 at 05:51

2

Thank you, this seems to be the only answer that gives `char` and its inverse in the case of unicode correctly – Munyari Aug 13 '16 at 16:25
important context [here](https://bugs.ruby-lang.org/issues/17405) and [here](https://github.com/ruby/irb/issues/43) if anyone comes this way. – New Alexandria Sep 24 '22 at 21:00

score 3 · Answer 6 · answered Nov 21 '08 at 13:45

3

Additionally, if you have the char in a string and you want to decode it without a loop:

puts 'Az'[0]
=> 65
puts 'Az'[1]
=> 122

answered Nov 21 '08 at 13:45

Kent Fredric

56,416
14
107
150

score 2 · Answer 7 · answered Mar 10 '14 at 19:38

2

You can have these:

65.chr.ord
'a'.ord.chr

answered Mar 10 '14 at 19:38

Eduardo Santana

5,780
3
19
21

score 2 · Answer 8 · answered May 24 '14 at 21:51

2

If you don't mind pulling the values out of an array, you can use "A".bytes

answered May 24 '14 at 21:51

Clark

416
3
10

score 2 · Answer 9 · answered Nov 21 '08 at 13:32

2

How about

puts ?A

answered Nov 21 '08 at 13:32

GregD

6,860
5
34
61

hantscolin · Answer 10 · 2015-04-14T07:42:01.190

0

I'm writing code for 1.8.6 and 1.9.3 and I couldn't get any of these solutions to work in both environments :(

However, I came across another solution: http://smajnr.net/2009/12/ruby-1-8-nomethoderror-undefined-method-ord-for-string.html

That didn't work for me either but I adapted it for my use:

unless "".respond_to?(:ord)
  class Fixnum
    def ord
      return self
    end
  end
end

Having done that, then the following will work in both environments

'A'[0].ord

edited Apr 14 '15 at 07:42

answered Apr 13 '15 at 15:11

hantscolin

1
2

When user18096 wrote their answer, `"A".unpack("C")[0]`, that was targeting Ruby 1.8.1, Ruby 1.8.7 and Ruby 1.9.2. Does it fail in your environment? What kind of failure? – RJHunter Apr 13 '15 at 22:32
Hi RJHunter, I am trying to convert a specific character in a string to its number value. The following code works in 1.9.3 but not 1.8.6. `self.status = tagAccountString[4].unpack('C')[0]` In 1.8.6 I get `Exception undefined method `unpack' for 0:Fixnum processing main buffered tag data - exit` The following code works (with my proposed solution) in both environments `self.status = tagAccountString[4].ord` Any advice (e.g. a better solution) is more than welcome – hantscolin Apr 15 '15 at 10:19
`tagAccountString[4]` returns a String in newer Rubies but used to return a Fixnum in Ruby 1.8. That's why you saw the error, `undefined method unpack for 0:Fixnum`. You could use `status = tagAccountString[4,1].unpack('C')[0]` or even `status, = tagAccountString.unpack('xxxxC')` if you always want to ignore four characters and convert the next one. – RJHunter Apr 18 '15 at 05:25
Thanks RJHunter for the explanation and alternative solutions. However, as "my" solution is more readable and reusable, I'll stick with that (unless there is a good reason not too?) – hantscolin Apr 22 '15 at 13:34

What's the opposite of chr() in Ruby?

10 Answers10

Linked