I want to find unique elements from an array of arrays by the first element in the inner arrays.
for example
a = [[1,2],[2,3],[1,5]
I want something like
[[1,2],[2,3]]
Add a comment
|
2 Answers
The uniq method takes a block:
uniq_a = a.uniq(&:first)
Or if you want to do it in-place:
a.uniq!(&:first)
For example:
>> a = [[1,2],[2,3],[1,5]]
=> [[1, 2], [2, 3], [1, 5]]
>> a.uniq(&:first)
=> [[1, 2], [2, 3]]
>> a
=> [[1, 2], [2, 3], [1, 5]]
Or
>> a = [[1,2],[2,3],[1,5]]
=> [[1, 2], [2, 3], [1, 5]]
>> a.uniq!(&:first)
=> [[1, 2], [2, 3]]
>> a
=> [[1, 2], [2, 3]]
If you're stuck back in 1.8.7 land where uniq doesn't take a block, then you can do it this way:
a.group_by(&:first).values.map(&:first)
For example:
>> a = [[1,2],[2,3],[1,5]]
=> [[1, 2], [2, 3], [1, 5]]
>> a.group_by(&:first).values.map(&:first)
=> [[1, 2], [2, 3]]
Thanks for the extra prodding Jin.
2 Comments
Jin
Note: I was trying this out in 1.8.7 and wondered why it wasn't working. Found out that passing a block to
uniq was added in Ruby 1.9.2, as stated here. rbjl.net/….
mu is too short
@Jin: Right, I added a 1.8.7 compatible version for you :)
Here's a ruby 1.8.7 solution
irb> [[1,2],[2,3],[1,5]].inject([]) { |memo,x| memo << x unless memo.detect { |item| item.first == x.first }; memo }
=> [[1, 2], [2, 3]]
You could also take the shorthand over a hash and be a bit lazy and take the last item instead
irb> [[1,2],[2,3],[1,5]].inject({}) { |memo,x| memo[x.first] = x; memo }.map { |x| x.last }
=> [[1, 5], [2, 3]]
1 Comment
mu is too short
Something like
[[1,2],[2,3],[1,5]].inject({}) { |memo,x| memo[x.first] ||= x; memo }.values would be another option (since we're only dealing with numbers in the arrays) and would probably faster than my fancy 1.8.7 version (not that the speed difference would matter much of course).