Difference Between sort and sort_by in Ruby

Ruby has two (and probably more) methods to easily sort collection: sort and sort_by. The distinction between the two methods is not clear while reading the documentation, therefore I will try to sum up the main points.

sort

sort compares two elements using the <=> operator or an optional code block.

If called with a block, it needs two parameters, noted a and b in the official documentation, which are compared and return -1 if the first operand (a) is less than the second (b), 0 if they are equal and 1 if the first operand is greater than the second. The return values shuffle up and down the elements until they are in the specified order.

[1, 5, 6, 2, 9, 2].sort { |a, b| b <=> a }  # => [9, 6, 5, 2, 2, 1]

sort_by

In contrast, sort_by needs to be called passing in a block in order to return the same collection object it was called upon (passing in nothing will return an enumerator).

The block itself needs only one parameter (a second parameter would return nil for each iteration). Inside the block, the argument needs to be manipulated in such a way as to make it possible to sort it. For example:

%w[rangers power].sort_by { |string| string.length }  # => ["power", "rangers"]

Here, string is rendered sortable by characters length thanks to the length method ¹ .

The other difference between sort and sort_by is that the later implements a Schwartzian transformation.

Simply explained, it assigns (or maps) a sort key to each element of the collection it is called upon, and sort thoses keys and mapped back onto the original values (well, that wasn’t simply explained after all).

sort_by is fairly expensive when used on small collection and sort should be prefered for simpler sorting.