Using Ruby 2.3:
In example 1, the string key \"a\"
is automatically converted to a symbol, whereas with example 2, it stays a string.
Example 1
According to Ruby documentation:
Blockquote Symbol objects represent names and some strings inside the Ruby interpreter. They are generated using the :name and :"string" literals syntax, and by the various to_sym methods. [...]
This means that running:
$ ruby -e ruby -e "h = {key: \"value\"}; puts h"
$ ruby -e ruby -e "h = {:key => \"value\"}; puts h"
$ ruby -e ruby -e "h = {\"key\": \"value\"}; puts h"
$ ruby -e ruby -e "h = {:\"key\" => \"value\"}; puts h"
$ ruby -e ruby -e "h = {\"#{:key}\": \"value\"}; puts h"
Will produce the same result:
$ {:key=>"value"}
In Ruby 2.3(.0), these are all the same:
{:"a" => 1}
{"a": 1},
{:a => 1}
{a: 1}
They all translate to the same thing: a
is a symbol in all these cases.
{"a"=>1}
is different: a
is a string in this case.
It's because of the new hash syntax introduced with ruby 1.9. The syntax with colon works with symbol keys only. It's called a "symbol to object" hash and it's only syntactic sugar for the most common style of hashes out there. Another point for me, it's closer to the javascript object notation.
If I have mixed key types then I prefer the old style (hash-rocket syntax), but that's up to you. Mixing the two style looks ugly to me.