Ruby 2.5.0 was recently released.
Earlier to Ruby 1.9 Hash#fetch method would raise IndexError for an invalid key.
An IndexError is raised when the given index value is invalid.
Ruby
irb(main):001:0> hash = { 'foo' => 'bar' }
irb(main):002:0> hash.fetch('foo')
IndexError: key not found
from (irb):2:in `fetch'
from (irb):2
from :0
KeyError class was introduced in Ruby 1.9 and it inherits from IndexError class. It is raised when the specified key is not found.
Ruby
>> hash = {foo: :bar}
=> {:foo=>:bar}
>> hash.fetch(:baz)
KeyError: key not found: :baz
from (irb):2:in `fetch'
from (irb):2
from /home/atul/.rvm/rubies/ruby-2.4.4/bin/irb:11:in `<main>'
It was proposed that having KeyError#receiver and KeyError#name
(or KeyError#key) like NameError, will help to debug and find typo on
did_you_mean gem.
Ruby
begin
h = {foo: 1, bar: 2}
h.fetch(:bax)
rescue KeyError => e
p e.name #=> :bax
p e.receiver #=> {foo: 1, bar: 2} equal `h`
end
There was a discussion about what method name is best between name and key.
Benoit Daloze suggested that name is more consistent with other error
classes such as NameError#name and NoMethodError#name. Having key as
method name was approved.
Ruby 2.5.0
Ruby 2.5.0 provides KeyError#receiver and KeyError#key as methods to return details of receiver object and key which was not found.
Ruby
begin
h = { foo: 1, bar: 2}
h.fetch(:bax)
rescue KeyError => e
p e.key #=> :bax
p e.receiver #=> {foo: 1, bar: 2}
end
I hope these methods are useful in your future debugging attempts.