Looking for a functional collapse/group-by of this data in ruby

Question

Can anyone come up with a functional (no mutating variables) way in ruby to do the following?

Given this sorted data:

data = [{1 => "A"},
        {2 => "B"},
        {3 => "B"},
        {4 => "C"},
        {5 => "D"},
        {6 => "D"},
        {7 => "D"}]

Yield this

[{"start" => 1, "end" => 1, "value" => "A"},
 {"start" => 2, "end" => 3, "value" => "B"},
 {"start" => 4, "end" => 4, "value" => "C"},
 {"start" => 5, "end" => 7, "value" => "D"}]

Answer 1

# v = data.map{|d| d.values[0]}
# k = data.map{|d| d.keys[0]}

k,v = [:keys, :values].map{|m| data.map{|d| d.send(m)[0]}}

seq = v.sort.uniq.map{|x| {:start => k[v.index(x)], :stop => k[v.rindex(x)], :value => x} }
seq.each do |p|
  p p
end

Answer 2

You need the abstraction "group consecutive elements by a given criterion", that's Enumerable#chunk:

data.chunk { |h| h.values.first }.map do |value, hs| 
  {"start" => hs.first.keys[0], "end" => hs.last.keys[0], "value" => value}
end

But note that this solution required some ugly gymnastics on the hash/array transformations. That's because the original data structure is not quite right, what sense does it make to have hashes with a single element? a simple pair will do. If you do that transformation on the data, the resulting code is easier to understand:

data.map(&:first).chunk { |index, char| char }.map do |char, pairs| 
  {"start" => pairs.first[0], "end" => pairs.last[0], "value" => char}
end