CODEWITHSUNDEEP
rubystringchunking
MiniQuark
MiniQuark

Reputation: 48495

What is the best way to chop a string into chunks of a given length in Ruby?

I have been looking for an elegant and efficient way to chunk a string into substrings of a given length in Ruby.

So far, the best I could come up with is this:

def chunk(string, size)
  (0..(string.length-1)/size).map{|i|string[i*size,size]}
end

>> chunk("abcdef",3)
=> ["abc", "def"]
>> chunk("abcde",3)
=> ["abc", "de"]
>> chunk("abc",3)
=> ["abc"]
>> chunk("ab",3)
=> ["ab"]
>> chunk("",3)
=> []

You might want chunk("", n) to return [""] instead of []. If so, just add this as the first line of the method:

return [""] if string.empty?

Would you recommend any better solution?

Edit

Thanks to Jeremy Ruten for this elegant and efficient solution: [edit: NOT efficient!]

def chunk(string, size)
    string.scan(/.{1,#{size}}/)
end

Edit

The string.scan solution takes about 60 seconds to chop 512k into 1k chunks 10000 times, compared with the original slice-based solution which only takes 2.4 seconds.

Upvotes: 103

Views: 43730

Answers (10)

Jason
Jason

Reputation: 2371

Here is another way to do it:

"abcdefghijklmnopqrstuvwxyz".chars.to_a.each_slice(3).to_a.map {|s| s.to_s }

Or,

"abcdefghijklmnopqrstuvwxyz".chars.each_slice(3).map(&:join)

Either:

=> ["abc", "def", "ghi", "jkl", "mno", "pqr", "stu", "vwx", "yz"]

Upvotes: 25

akim
akim

Reputation: 8769

I personally followed the idea of user8556428, to avoid the costly intermediate values that most proposals introduce, and to avoid modifying the input string. And I want to be able to use it as a generator (for instance to use s.each_slice.with_index).

My use case is really about bytes, not characters. In the case of character-size, strscan is a great solution.

class String
    # Slices of fixed byte-length.  May cut multi-byte characters.
    def each_slice(n = 1000, &block)
        return if self.empty?

        if block_given?
            last = (self.length - 1) / n
            (0 .. last).each do |i|
                yield self.slice(i * n, n)
            end
        else
            enum_for(__method__, n)
        end
    end
end


p "abcdef".each_slice(3).to_a # => ["abc", "def"]   
p "abcde".each_slice(3).to_a  # => ["abc", "de"]    
p "abc".each_slice(3).to_a    # => ["abc"]          
p "ab".each_slice(3).to_a     # => ["ab"]           
p "".each_slice(3).to_a       # => []

Upvotes: 0

Vyacheslav
Vyacheslav

Reputation: 27221

Just text.scan(/.{1,4}/m) resolves the problem

Upvotes: 0

prcu
prcu

Reputation: 1013

Here is another one solution for slightly different case, when processing large strings and there is no need to store all chunks at a time. In this way it stores single chunk at a time and performs much faster than slicing strings:

io = StringIO.new(string)
until io.eof?
  chunk = io.read(chunk_size)
  do_something(chunk)
end

Upvotes: 7

Per Wigren
Per Wigren

Reputation: 61

I made a little test that chops about 593MB data into 18991 32KB pieces. Your slice+map version ran for at least 15 minutes using 100% CPU before I pressed ctrl+C. This version using String#unpack finished in 3.6 seconds:

def chunk(string, size)
  string.unpack("a#{size}" * (string.size/size.to_f).ceil)
end

Upvotes: 6

kirkytullins
kirkytullins

Reputation: 152

A better solution which takes into account the last part of the string which could be less than the chunk size:

def chunk(inStr, sz)  
  return [inStr] if inStr.length < sz  
  m = inStr.length % sz # this is the last part of the string
  partial = (inStr.length / sz).times.collect { |i| inStr[i * sz, sz] }
  partial << inStr[-m..-1] if (m % sz != 0) # add the last part 
  partial
end

Upvotes: 1

davispuh
davispuh

Reputation: 1433

I think this is the most efficient solution if you know your string is a multiple of chunk size

def chunk(string, size)
    (string.length / size).times.collect { |i| string[i * size, size] }
end

and for parts

def parts(string, count)
    size = string.length / count
    count.times.collect { |i| string[i * size, size] }
end

Upvotes: 6

Paige Ruten
Paige Ruten

Reputation: 176753

Use String#scan:

>> 'abcdefghijklmnopqrstuvwxyz'.scan(/.{4}/)
=> ["abcd", "efgh", "ijkl", "mnop", "qrst", "uvwx"]
>> 'abcdefghijklmnopqrstuvwxyz'.scan(/.{1,4}/)
=> ["abcd", "efgh", "ijkl", "mnop", "qrst", "uvwx", "yz"]
>> 'abcdefghijklmnopqrstuvwxyz'.scan(/.{1,3}/)
=> ["abc", "def", "ghi", "jkl", "mno", "pqr", "stu", "vwx", "yz"]

Upvotes: 177

Chuck
Chuck

Reputation: 237100

test.split(/(...)/).reject {|v| v.empty?}

The reject is necessary because it otherwise includes the blank space between sets. My regex-fu isn't quite up to seeing how to fix that right off the top of my head.

Upvotes: 1

Charlie Martin
Charlie Martin

Reputation: 112406

Are there some other constraints you have in mind? Otherwise I'd be awfully tempted to do something simple like

[0..10].each {
   str[(i*w),w]
}

Upvotes: 0

Related Questions