bcobb

Gratuitous Ruby: Counting Item Frequency

Dec 8, 2014

Tonight on Twitter, @jessitron posted the following:

Counting in #Ruby: ["a","b","a"].inject(Hash.new(0)) { |m,o| m[o] += 1; m } => {"a"=>2, "b"=>1} @mattruzicka #STLRuby

— Jessica Kerr (@jessitron) December 9, 2014

The other week, I needed to do exactly this, and came up with a slightly different approach. Here’s my (gently elaborated) version of Jessica’s snippet:

%w(a b a).reduce(Hash.new) do |map, item|
  map.merge(item => 1) do |key, sum, increment|
    sum + increment
  end
end

I don’t know that it’s better or worse, but it has two properties I like:

1. The body of reduce is immutable.
2. It utilizes merge’s ability to take a block to resolve conflicting updates.¹

The last point is the linchpin of the block. When we merge in a new item key into map, its value is set to 1. If we encounter that key again, merge sets the value key to the result of its block. The block itself takes three arguments: the conflicting key, the existing value of that key, and the value we attempted to merge. For our purposes, the key isn’t necessary; but notice that its existing value is simply the number of item keys we’ve tried to merge in the past (starting with 1), and the value we’re attempting to merge is the value by which we increment the count.

Anyway, there’s not really a point to this post, other than that it can be fun to fart around with Enumerable on a sleety Monday night.

Postscript, 2017-07-29

These days, I’d almost certainly do what Jessica wrote in her initial toot, with one small change:

%(a b a).each_with_object(Hash.new(0)) do |i, h|
  h[i] += 1
end

Why is this? Well: these days, I actually understand what each_with_object does! In 2014, that was not the case.