How does count method works in Ruby? - ruby

Question

I have a hard time understanding the following code segment from the Ruby docs:
a = "hello world"
a.count "lo" #=> 5
a.count "lo", "o" #=> 2
a.count "hello", "^l" #=> 4
a.count "ej-m" #=> 4
"hello^world".count "\\^aeiou" #=> 4
"hello-world".count "a\\-eo" #=> 4
especially this code a.count "ej-m". Can anyone please explain how it works?

Answer 1

Just imagine the "pattern" strings as wrapped by [ and ] from regex syntax, that are matched against each character.
So, if we break a = "hello world" into characters:
[1] pry(main)> a = "hello world"
=> "hello world"
[2] pry(main)> a.split('')
=> ["h", "e", "l", "l", "o", " ", "w", "o", "r", "l", "d"]
And convert "ej-m" to regex wrapped with [ and ] we get /[ej-m]/ - which means either 'e' or any character from 'j' to 'm'(including both):
[3] pry(main)> a.split('').select{|c| c=~ /[ej-m]/}
=> ["e", "l", "l", "l"]
We got 4 matches - which is also the result you get. Essensially a.count "ej-m" is equivalent to:
[4] pry(main)> a.split('').count{|c| c=~ /[ej-m]/}
=> 4
Multiple arguments to the method are just and between the matches:
[5] pry(main)> a.split('').count{|c| c =~ /[hello]/ and c =~ /[^l]/}
=> 4

Answer 2

The sequence c1-c2 means all characters between c1 and c2.
So you are providing a range, basically it counts which characters are in that range (>= c1 && <= c2)
i.e:
a = "hello world"
a.count "a-z"
=> 10
a.count "o-w"
=> 4 #(o, o, r, w)
a.count "e-l"
=> 5 #(h, e, l, l, l)

Answer 3

We find that
"hello world".count("ej-m") #=> 4 (_ell_____l_)
Examine the doc for String#count carefully.
Here is how count might be implemented to deal with patterns that closely resemble the pattern "ej-m".
def count_letters(str, pattern)
idx = pattern[1..-2].index('-')
if idx
idx += 1
before, after = pattern[idx-1], pattern[idx+1]
pattern[idx-1..idx+1] = (before..after).to_a.join
end
str.each_char.sum { |c| pattern.include?(c) ? 1 : 0 }
end
count_letters(str, pattern) #=> 4 (_ell_____l_)
However, String#count must also do the following.
Allow for multiple ranges in the pattern
"hello1world".count("e0-9j-mv-x") #=> 6 (_ell__1_w__l_)
If the pattern begins with the character '^'count the number of characters that do not match the remainder of the pattern
"hello world".count("^ej-m") #=> 7 (h___o*wor_d) * = space to count
"hello^world".count("e^j-m") #=> 5 (_ell_^___l_)
"hello world".count("\^ej-m") #=> 7 (h___o*wor_d) * = space to count
Note that escaping '^' at the beginning of the string makes no difference.
Match a hyphen
"hello-world".count("ej-m-") #=> 5 (_ell_-___l_)
"hello-world".count("-ej-m") #=> 5 (_ell_-___l_)
"hello-world".count("ej\-m") #=> 4 (_ell____l_)
Note that escaping a hyphen that is not the first or last character of the pattern makes no difference.
Match a backslash
'hello\world'.count("ej-m\\") #=> 5 (_ell_\___l_)
'hello\world'.count("\\ej-m") #=> 4 (_ell____l_)
Note that a backslash at the beginning of a string is disregarded.
Some of the above results (Ruby v2.4) do not seem to be consistent with the documentation.