[ruby-core:90847] [Ruby trunk Feature#15492] Let #dig take a "default value" block like Hash#fetch does

[ruby-core:90847] [Ruby trunk Feature#15492] Let #dig take a "default value" block like Hash#fetch does

[tyler]

Issue #15492 has been reported by TylerRick (Tyler Rick).

----------------------------------------
Feature #15492: Let #dig take a "default value" block like Hash#fetch does
https://bugs.ruby-lang.org/issues/15492

* Author: TylerRick (Tyler Rick)
* Status: Open
* Priority: Normal
* Assignee: 
* Target version: 
----------------------------------------
[fetch](https://ruby-doc.org/core-2.6/Hash.html#method-i-fetch) provides multiple ways to handle the case where a key can't be found:

> If the key can't be found, there are several options: With no other arguments, it will raise a KeyError exception; if `default` is given, then that will be returned; if the optional code block is specified, then that will be run and its result returned.

~~~ ruby
{a: 'a'}.fetch(:d) { 'default' }
#=> "default"

{a: 'a'}.fetch(:d) {|key| key }
#=> :d
~~~

[dig](https://ruby-doc.org/core-2.6/Hash.html#method-i-dig) obviously can't let you specify a `default` value as a positional argument like `fetch` does, but couldn't it at least let you specify a default value by passing a block?

The fact that it currently just silently *ignores* the block that you pass to `dig` could be misleading, as one might assume (as I did at first) that it's going to return that in case any of the key lookups fail. 

Current/desired behavior:

~~~ ruby
{a: {b: {c: 'c'}}}.dig(:a, :b, :d) {|key| key }
#=> nil
# wish it returned :d

{a: {b: {c: 'c'}}}.dig(:a, :b, :d) {|key| 'default' }
#=> nil
# wish it returned 'default'
~~~

There isn't currently a nice way to do this (that I can think of). You could mix `dig` and `fetch` (or chain a bunch of `fetch`s), but that loses the simple elegance of `dig` and looks awful:

~~~ ruby
object = {a: {b: {c: 'c'}}}
(object.dig(:a, :b) || {}).fetch(:d) { 'default' }
#=> "default"
~~~

Note, of course, that if we added default-block behavior to `dig`, its behavior would be slightly different: it would return the result of the default block if _any_ intermediate step were `nil`, not just if the _last_ lookup were nil.

~~~ ruby
object = {a: {b: {c: 'c'}}}
object.dig(:x, :y, :z) { 'default' }
#=> "default"
~~~


## Example use case

Sometimes you start out with a simple Hash but over time you may end up moving keys into sub-hashes.

You might have started out with something like

~~~ ruby
config.fetch(:something) { 'default' }
~~~

But after you move `:something` under `:settings`, you have to use `dig` instead of `fetch`:

~~~ ruby
config.dig(:settings, :something)
~~~

The problem is, how do you keep the default value now that it's in a sub-hash? Can't just use `|| 'default'` because the stored value might be `false`.

Currently I'm working around this with an explicit `nil?` check...

~~~ ruby
value = config.dig(:settings, :something)
value.nil? ? 'default' : value
~~~

(Even *that* may not be good enough if one wanted to distinguish between a `nil` value _stored_ in the object and the case where the key can't be found (a "cache miss"). Having a default block would let you distinguish between a missing key and a nil value stored at the key, in case that distinction were important...)



-- 
https://bugs.ruby-lang.org/