Well you are doing great so far.

Piping is just syntax sugar, using it is not refactoring. What you have is perfectly readable so I think it’s fine.

Not always error-handling exactly - Map.fetch is really useful when it’s important to your code whether the key is present at all and the value could potentially be nil.

For instance, if you need to distinguish:

without_baz = %{foo: "bar"}
# from
nil_baz = %{foo: "bar", baz: nil}

then you’ll get different results with different functions in Map:

# square-brackets return nil if the key isn't present
nil_baz[:baz] == without_baz[:baz]

# Map.get can return a default
Map.get(without_baz, :baz, "default value") # => "default value"
# vs
Map.get(nil_baz, :baz, "default value") # => nil

# But if there's no "safe" default that won't appear in the source, Map.fetch is better:
Map.fetch(without_baz, :baz) # => :error
# vs
Map.fetch(nil_baz, :baz) # => {:ok, nil}