What's a convenient way to get a list of all Hash keys (with nesting) separated by dots?
Given I have a hash:
{ level1: { level21: { level31: 'val1',
level32: 'val2' },
level22: 'val3' }
}
Desired output (array of strings) which represents all key paths in a hash:
level1.level21.level31
level1.level21.level32
level1.level22
My current solution:
class HashKeysDumper
def self.dump(hash)
hash.map do |k, v|
if v.is_a? Hash
keys = dump(v)
keys.map { |k1| [k, k1].join('.') }
else
k.to_s
end
end.flatten
end
end
It also available as gist (with specs).
Well, it depends on what you mean by cleaner, but here's a smaller version that…
Will work on subclasses Hashes or Hash-alikes
Extends Hash, making it look cleaner in your code.
class Hash
def keydump
map{|k,v|v.keydump.map{|a|"#{k}.#{a}"} rescue k.to_s}.flatten
end
end
results:
{ level1: { level21: { level31: 'val1',
level32: 'val2' },
level22: 'val3' }
}.keydump
=> ["level1.level21.level31", "level1.level21.level32", "level1.level22"]
Here is my vision of this:
h = { 'level1' => { 'level2' => { 'level31' => 'val1', 'level32' => 'val2' } } }
class Hash
def nested_keys
self.inject([]) { |f, (k,v)| f += [k, v.is_a?(Hash) ? v.nested_keys : []] }.flatten
end
end
keys = h.nested_keys
p keys
#=> ["level1", "level2", "level31", "level32"]
k1, k2 = keys.shift, keys.shift
puts [k1, k2, keys.shift].join('.')
#=> level1.level2.level31
puts [k1, k2, keys.shift].join('.')
#=> level1.level2.level32
Here is a Working Demo
I just committed some code to RubyTree that adds from_hash() which would allow you to do this:
require 'rubytree'
Tree::TreeNode.from_hash(hash).each_leaf.map{|n| "#{n.name}.#{n.parentage.map(&:name).reverse.join('.')}" }
=> ["level1.level21.level31", "level1.level21.level32", "level1.level22"]
Aside from the gem require, it's a one-liner :)
Related
Given I have this hash:
h = { a: 'a', b: 'b', c: { d: 'd', e: 'e'} }
And I convert to OpenStruct:
o = OpenStruct.new(h)
=> #<OpenStruct a="a", b="b", c={:d=>"d", :e=>"e"}>
o.a
=> "a"
o.b
=> "b"
o.c
=> {:d=>"d", :e=>"e"}
2.1.2 :006 > o.c.d
NoMethodError: undefined method `d' for {:d=>"d", :e=>"e"}:Hash
I want all the nested keys to be methods as well. So I can access d as such:
o.c.d
=> "d"
How can I achieve this?
You can monkey-patch the Hash class
class Hash
def to_o
JSON.parse to_json, object_class: OpenStruct
end
end
then you can say
h = { a: 'a', b: 'b', c: { d: 'd', e: 'e'} }
o = h.to_o
o.c.d # => 'd'
See Convert a complex nested hash to an object.
I came up with this solution:
h = { a: 'a', b: 'b', c: { d: 'd', e: 'e'} }
json = h.to_json
=> "{\"a\":\"a\",\"b\":\"b\",\"c\":{\"d\":\"d\",\"e\":\"e\"}}"
object = JSON.parse(json, object_class:OpenStruct)
object.c.d
=> "d"
So for this to work, I had to do an extra step: convert it to json.
personally I use the recursive-open-struct gem - it's then as simple as RecursiveOpenStruct.new(<nested_hash>)
But for the sake of recursion practice, I'll show you a fresh solution:
require 'ostruct'
def to_recursive_ostruct(hash)
result = hash.each_with_object({}) do |(key, val), memo|
memo[key] = val.is_a?(Hash) ? to_recursive_ostruct(val) : val
end
OpenStruct.new(result)
end
puts to_recursive_ostruct(a: { b: 1}).a.b
# => 1
edit
Weihang Jian showed a slight improvement to this here https://stackoverflow.com/a/69311716/2981429
def to_recursive_ostruct(hash)
hash.each_with_object(OpenStruct.new) do |(key, val), memo|
memo[key] = val.is_a?(Hash) ? to_recursive_ostruct(val) : val
end
end
Also see https://stackoverflow.com/a/63264908/2981429 which shows how to handle arrays
note
the reason this is better than the JSON-based solutions is because you can lose some data when you convert to JSON. For example if you convert a Time object to JSON and then parse it, it will be a string. There are many other examples of this:
class Foo; end
JSON.parse({obj: Foo.new}.to_json)["obj"]
# => "#<Foo:0x00007fc8720198b0>"
yeah ... not super useful. You've completely lost your reference to the actual instance.
Here's a recursive solution that avoids converting the hash to json:
def to_o(obj)
if obj.is_a?(Hash)
return OpenStruct.new(obj.map{ |key, val| [ key, to_o(val) ] }.to_h)
elsif obj.is_a?(Array)
return obj.map{ |o| to_o(o) }
else # Assumed to be a primitive value
return obj
end
end
My solution is cleaner and faster than #max-pleaner's.
I don't actually know why but I don't instance extra Hash objects:
def dot_access(hash)
hash.each_with_object(OpenStruct.new) do |(key, value), struct|
struct[key] = value.is_a?(Hash) ? dot_access(value) : value
end
end
Here is the benchmark for you reference:
require 'ostruct'
def dot_access(hash)
hash.each_with_object(OpenStruct.new) do |(key, value), struct|
struct[key] = value.is_a?(Hash) ? dot_access(value) : value
end
end
def to_recursive_ostruct(hash)
result = hash.each_with_object({}) do |(key, val), memo|
memo[key] = val.is_a?(Hash) ? to_recursive_ostruct(val) : val
end
OpenStruct.new(result)
end
require 'benchmark/ips'
Benchmark.ips do |x|
hash = { a: 1, b: 2, c: { d: 3 } }
x.report('dot_access') { dot_access(hash) }
x.report('to_recursive_ostruct') { to_recursive_ostruct(hash) }
end
Warming up --------------------------------------
dot_access 4.843k i/100ms
to_recursive_ostruct 5.218k i/100ms
Calculating -------------------------------------
dot_access 51.976k (± 5.0%) i/s - 261.522k in 5.044482s
to_recursive_ostruct 50.122k (± 4.6%) i/s - 250.464k in 5.008116s
My solution, based on max pleaner's answer and similar to Xavi's answer:
require 'ostruct'
def initialize_open_struct_deeply(value)
case value
when Hash
OpenStruct.new(value.transform_values { |hash_value| send __method__, hash_value })
when Array
value.map { |element| send __method__, element }
else
value
end
end
Here is one way to override the initializer so you can do OpenStruct.new({ a: "b", c: { d: "e", f: ["g", "h", "i"] }}).
Further, this class is included when you require 'json', so be sure to do this patch after the require.
class OpenStruct
def initialize(hash = nil)
#table = {}
if hash
hash.each_pair do |k, v|
self[k] = v.is_a?(Hash) ? OpenStruct.new(v) : v
end
end
end
def keys
#table.keys.map{|k| k.to_s}
end
end
Basing a conversion on OpenStruct works fine until it doesn't. For instance, none of the other answers here properly handle these simple hashes:
people = { person1: { display: { first: 'John' } } }
creds = { oauth: { trust: true }, basic: { trust: false } }
The method below works with those hashes, modifying the input hash rather than returning a new object.
def add_indifferent_access!(hash)
hash.each_pair do |k, v|
hash.instance_variable_set("##{k}", v.tap { |v| send(__method__, v) if v.is_a?(Hash) } )
hash.define_singleton_method(k, proc { hash.instance_variable_get("##{k}") } )
end
end
then
add_indifferent_access!(people)
people.person1.display.first # => 'John'
Or if your context calls for a more inline call structure:
creds.yield_self(&method(:add_indifferent_access!)).oauth.trust # => true
Alternatively, you could mix it in:
module HashExtension
def very_indifferent_access!
each_pair do |k, v|
instance_variable_set("##{k}", v.tap { |v| v.extend(HashExtension) && v.send(__method__) if v.is_a?(Hash) } )
define_singleton_method(k, proc { self.instance_variable_get("##{k}") } )
end
end
end
and apply to individual hashes:
favs = { song1: { title: 'John and Marsha', author: 'Stan Freberg' } }
favs.extend(HashExtension).very_indifferent_access!
favs.song1.title
Here is a variation for monkey-patching Hash, should you opt to do so:
class Hash
def with_very_indifferent_access!
each_pair do |k, v|
instance_variable_set("##{k}", v.tap { |v| v.send(__method__) if v.is_a?(Hash) } )
define_singleton_method(k, proc { instance_variable_get("##{k}") } )
end
end
end
# Note the omission of "v.extend(HashExtension)" vs. the mix-in variation.
Comments to other answers expressed a desire to retain class types. This solution accommodates that.
people = { person1: { created_at: Time.now } }
people.with_very_indifferent_access!
people.person1.created_at.class # => Time
Whatever solution you choose, I recommend testing with this hash:
people = { person1: { display: { first: 'John' } }, person2: { display: { last: 'Jingleheimer' } } }
If you are ok with monkey-patching the Hash class, you can do:
require 'ostruct'
module Structurizable
def each_pair(&block)
each do |k, v|
v = OpenStruct.new(v) if v.is_a? Hash
yield k, v
end
end
end
Hash.prepend Structurizable
people = { person1: { display: { first: 'John' } }, person2: { display: { last: 'Jingleheimer' } } }
puts OpenStruct.new(people).person1.display.first
Ideally, instead of pretending this, we should be able to use a Refinement, but for some reason I can't understand it didn't worked for the each_pair method (also, unfortunately Refinements are still pretty limited)
Suppose I have following hash or nested hash:
h = { :a1 => { :b1 => "c1" },
:a2 => { :b2 => "c2"},
:a3 => { :b3 => "c3"} }
I want to create a method that takes hash as a parameter and recursively convert all the keys (keys that are symbol eg. :a1) to String (eg. "a1"). So far I have come up with the following method which doesn't work and returns {"a1"=>{:b1=>"c1"}, "a2"=>{:b2=>"c2"}, "a3"=>{:b3=>"c3"}}.:
def stringify_all_keys(hash)
stringified_hash = {}
hash.each do |k, v|
stringified_hash[k.to_s] = v
if v.class == Hash
stringify_all_keys(stringified_hash[k.to_s])
end
end
stringified_hash
end
What am I doing wrong and how do a get all the keys converted to string like this:
{"a1"=>{"b1"=>"c1"}, "a2"=>{"b2"=>"c2"}, "a3"=>{"b3"=>"c3"}}
If you are using ActiveSupport already or are open to using it, then deep_stringify_keys is what you're looking for.
hash = { person: { name: 'Rob', age: '28' } }
hash.deep_stringify_keys
# => {"person"=>{"name"=>"Rob", "age"=>"28"}}
Quick'n'dirty if your values are basic objects like strings, numbers, etc:
require 'json'
JSON.parse(JSON.dump(hash))
Didn't test this, but looks about right:
def stringify_all_keys(hash)
stringified_hash = {}
hash.each do |k, v|
stringified_hash[k.to_s] = v.is_a?(Hash) ? stringify_all_keys(v) : v
end
stringified_hash
end
using plain ruby code, the below code could help.
you can monkey patched it to the ruby Hash, to use it like this my_hash.deeply_stringfy_keys
however, I do not recommend monkey batching ruby.
you can adjust the method to provide the deeply_strigify_keys! (bang) version of it.
in case you want to make a different method witch does not stringify recursively, or to control the level of stringifying then consider re-writing the below method logic so you can have it written better with considering the other variation mentioned above.
def deeply_stringify_keys(hash)
stringified_hash = {}
hash.each do |k, v|
if v.is_a?(Hash)
stringified_hash[k.to_s] = deeply_stringify_keys(v)
elsif v.is_a?(Array)
stringified_hash[k.to_s] = v.map {|i| i.is_a?(Hash)? deeply_stringify_keys(i) : i}
else
stringified_hash[k.to_s] = v
end
end
stringified_hash
end
h = {
data: {
user: {
value: "John Doe"
}
}
}
To assign value to the nested hash, we can use
h[:data][:user][:value] = "Bob"
However if any part in the middle is missing, it will cause error.
Something like
h.dig(:data, :user, :value) = "Bob"
won't work, since there's no Hash#dig= available yet.
To safely assign value, we can do
h.dig(:data, :user)&.[]=(:value, "Bob") # or equivalently
h.dig(:data, :user)&.store(:value, "Bob")
But is there better way to do that?
It's not without its caveats (and doesn't work if you're receiving the hash from elsewhere), but a common solution is this:
hash = Hash.new {|h,k| h[k] = h.class.new(&h.default_proc) }
hash[:data][:user][:value] = "Bob"
p hash
# => { :data => { :user => { :value => "Bob" } } }
And building on #rellampec's answer, ones that does not throw errors:
def dig_set(obj, keys, value)
key = keys.first
if keys.length == 1
obj[key] = value
else
obj[key] = {} unless obj[key]
dig_set(obj[key], keys.slice(1..-1), value)
end
end
obj = {d: 'hey'}
dig_set(obj, [:a, :b, :c], 'val')
obj #=> {d: 'hey', a: {b: {c: 'val'}}}
interesting one:
def dig_set(obj, keys, value)
if keys.length == 1
obj[keys.first] = value
else
dig_set(obj[keys.first], keys.slice(1..-1), value)
end
end
will raise an exception anyways if there's no [] or []= methods.
I found a simple solution to set the value of a nested hash, even if a parent key is missing, even if the hash already exists. Given:
x = { gojira: { guitar: { joe: 'charvel' } } }
Suppose you wanted to include mario's drum to result in:
x = { gojira: { guitar: { joe: 'charvel' }, drum: { mario: 'tama' } } }
I ended up monkey-patching Hash:
class Hash
# ensures nested hash from keys, and sets final key to value
# keys: Array of Symbol|String
# value: any
def nested_set(keys, value)
raise "DEBUG: nested_set keys must be an Array" unless keys.is_a?(Array)
final_key = keys.pop
return unless valid_key?(final_key)
position = self
for key in keys
return unless valid_key?(key)
position[key] = {} unless position[key].is_a?(Hash)
position = position[key]
end
position[final_key] = value
end
private
# returns true if key is valid
def valid_key?(key)
return true if key.is_a?(Symbol) || key.is_a?(String)
raise "DEBUG: nested_set invalid key: #{key} (#{key.class})"
end
end
usage:
x.nested_set([:instrument, :drum, :mario], 'tama')
usage for your example:
h.nested_set([:data, :user, :value], 'Bob')
any caveats i missed? any better way to write the code without sacrificing readability?
Searching for an answer to a similar question I developmentally stumbled upon an interface similar to #niels-kristian's answer, but wanted to also support a namespace definition parameter, like an xpath.
def deep_merge(memo, source)
# From: http://www.ruby-forum.com/topic/142809
# Author: Stefan Rusterholz
merger = proc { |key, v1, v2| Hash === v1 && Hash === v2 ? v1.merge(v2, &merger) : v2 }
memo.merge!(source, &merger)
end
# Like Hash#dig, but for setting a value at an xpath
def bury(memo, xpath, value, delimiter=%r{\.})
xpath = xpath.split(delimiter) if xpath.respond_to?(:split)
xpath.map!{|x|x.to_s.to_sym}.push(value)
deep_merge(memo, xpath.reverse.inject { |memo, field| {field.to_sym => memo} })
end
Nested hashes are sort of like xpaths, and the opposite of dig is bury.
irb(main):014:0> memo = {:test=>"value"}
=> {:test=>"value"}
irb(main):015:0> bury(memo, 'test.this.long.path', 'value')
=> {:test=>{:this=>{:long=>{:path=>"value"}}}}
irb(main):016:0> bury(memo, [:test, 'this', 2, 4.0], 'value')
=> {:test=>{:this=>{:long=>{:path=>"value"}, :"2"=>{:"4.0"=>"value"}}}}
irb(main):017:0> bury(memo, 'test.this.long.path.even.longer', 'value')
=> {:test=>{:this=>{:long=>{:path=>{:even=>{:longer=>"value"}}}, :"2"=>{:"4.0"=>"value"}}}}
irb(main):018:0> bury(memo, 'test.this.long.other.even.longer', 'other')
=> {:test=>{:this=>{:long=>{:path=>{:even=>{:longer=>"value"}}, :other=>{:even=>{:longer=>"other"}}}, :"2"=>{:"4.0"=>"value"}}}}
A more ruby-helper-like version of #niels-kristian answer
You can use it like:
a = {}
a.bury!([:a, :b], "foo")
a # => {:a => { :b => "foo" }}
class Hash
def bury!(keys, value)
key = keys.first
if keys.length == 1
self[key] = value
else
self[key] = {} unless self[key]
self[key].bury!(keys.slice(1..-1), value)
end
self
end
end
I have an OpenStruct that is nested with many other OpenStructs. What's the best way to deeply convert them all to JSON?
Ideally:
x = OpenStruct.new
x.y = OpenStruct.new
x.y.z = OpenStruct.new
z = 'hello'
x.to_json
// {y: z: 'hello'}
Reality
{ <OpenStruct= ....> }
There is no default methods to accomplish such task because the built-in #to_hash returns the Hash representation but it doesn't deep converts the values.
If a value is an OpenStruct, it's returned as such and it's not converted into an Hash.
However, this is not that complicated to solve. You can create a method that traverses each key/value in an OpenStruct instance (e.g. using each_pair), recursively descends into the nested OpenStructs if the value is an OpenStruct and returns an Hash of just Ruby basic types.
Such Hash can then easily be serialized using either .to_json or JSON.dump(hash).
This is a very quick example, with an update from #Yuval Rimar for arrays of OpenStructs:
def openstruct_to_hash(object, hash = {})
case object
when OpenStruct then
object.each_pair do |key, value|
hash[key] = openstruct_to_hash(value)
end
hash
when Array then
object.map { |v| openstruct_to_hash(v) }
else object
end
end
openstruct_to_hash(OpenStruct.new(foo: 1, bar: OpenStruct.new(baz: 2)))
# => {:foo=>1, :bar=>{:baz=>2}}
Fixes to above solution to handle arrays
def open_struct_to_hash(object, hash = {})
object.each_pair do |key, value|
hash[key] = case value
when OpenStruct then open_struct_to_hash(value)
when Array then value.map { |v| open_struct_to_hash(v) }
else value
end
end
hash
end
Here's yet another approach, modified from lancegatlin's answer. Also adding the method to the OpenStruct class itself.
class OpenStruct
def deep_to_h
each_pair.map do |key, value|
[
key,
case value
when OpenStruct then value.deep_to_h
when Array then value.map {|el| el === OpenStruct ? el.deep_to_h : el}
else value
end
]
end.to_h
end
in initializers/open_struct.rb:
require 'ostruct'
# Because #table is a instance variable of OpenStruct and Object#as_json returns Hash of instance variables.
class OpenStruct
def as_json(options = nil)
#table.as_json(options)
end
end
Usage:
OpenStruct.new({ a: { b: 123 } }).as_json
# Result
{
"a" => {
"b" => 123
}
}
Edit:
This seems to do almost the same thing (notice the keys are symbols instead of strings)
OpenStruct.new({ a: { b: 123 } }).marshal_dump
# Result
{
:a => {
:b => 123
}
}
Same function that can accept arrays as an input too
def openstruct_to_hash(object, hash = {})
case object
when OpenStruct then
object.each_pair do |key, value|
hash[key] = openstruct_to_hash(value)
end
hash
when Array then
object.map { |v| openstruct_to_hash(v) }
else object
end
end
None of the above worked for me with a copy and paste. Adding this solution:
require 'json'
class OpenStruct
def deep_to_h
each_pair.map do |key, value|
[
key,
case value
when OpenStruct then value.deep_to_h
when Array then value.map {|el| el.class == OpenStruct ? el.deep_to_h : el}
else value
end
]
end.to_h
end
end
json=<<HERE
{
"string": "fooval",
"string_array": [
"arrayval"
],
"int": 2,
"hash_array": [
{
"string": "barval",
"string2": "bazval"
},
{
"string": "barval2",
"string2": "bazval2"
}
]
}
HERE
os = JSON.parse(json, object_class: OpenStruct)
puts JSON.pretty_generate os.to_h
puts JSON.pretty_generate os.deep_to_h
I have a simple array:
arr = ["apples", "bananas", "coconuts", "watermelons"]
I also have a function f that will perform an operation on a single string input and return a value. This operation is very expensive, so I would like to memoize the results in the hash.
I know I can make the desired hash with something like this:
h = {}
arr.each { |a| h[a] = f(a) }
What I'd like to do is not have to initialize h, so that I can just write something like this:
h = arr.(???) { |a| a => f(a) }
Can that be done?
Say you have a function with a funtastic name: "f"
def f(fruit)
fruit + "!"
end
arr = ["apples", "bananas", "coconuts", "watermelons"]
h = Hash[ *arr.collect { |v| [ v, f(v) ] }.flatten ]
will give you:
{"watermelons"=>"watermelons!", "bananas"=>"bananas!", "apples"=>"apples!", "coconuts"=>"coconuts!"}
Updated:
As mentioned in the comments, Ruby 1.8.7 introduces a nicer syntax for this:
h = Hash[arr.collect { |v| [v, f(v)] }]
Did some quick, dirty benchmarks on some of the given answers. (These findings may not be exactly identical with yours based on Ruby version, weird caching, etc. but the general results will be similar.)
arr is a collection of ActiveRecord objects.
Benchmark.measure {
100000.times {
Hash[arr.map{ |a| [a.id, a] }]
}
}
Benchmark #real=0.860651, #cstime=0.0, #cutime=0.0, #stime=0.0, #utime=0.8500000000000005, #total=0.8500000000000005
Benchmark.measure {
100000.times {
h = Hash[arr.collect { |v| [v.id, v] }]
}
}
Benchmark #real=0.74612, #cstime=0.0, #cutime=0.0, #stime=0.010000000000000009, #utime=0.740000000000002, #total=0.750000000000002
Benchmark.measure {
100000.times {
hash = {}
arr.each { |a| hash[a.id] = a }
}
}
Benchmark #real=0.627355, #cstime=0.0, #cutime=0.0, #stime=0.010000000000000009, #utime=0.6199999999999974, #total=0.6299999999999975
Benchmark.measure {
100000.times {
arr.each_with_object({}) { |v, h| h[v.id] = v }
}
}
Benchmark #real=1.650568, #cstime=0.0, #cutime=0.0, #stime=0.12999999999999998, #utime=1.51, #total=1.64
In conclusion
Just because Ruby is expressive and dynamic, doesn't mean you should always go for the prettiest solution. The basic each loop was the fastest in creating a hash.
h = arr.each_with_object({}) { |v,h| h[v] = f(v) }
Ruby 2.6.0 enables a shorter syntax by passing a block to the to_h method:
arr.to_h { |a| [a, f(a)] }
This is what I would probably write:
h = Hash[arr.zip(arr.map(&method(:f)))]
Simple, clear, obvious, declarative. What more could you want?
I'm doing it like described in this great article http://robots.thoughtbot.com/iteration-as-an-anti-pattern#build-a-hash-from-an-array
array = ["apples", "bananas", "coconuts", "watermelons"]
hash = array.inject({}) { |h,fruit| h.merge(fruit => f(fruit)) }
More info about inject method: http://ruby-doc.org/core-2.0.0/Enumerable.html#method-i-inject
Another one, slightly clearer IMHO -
Hash[*array.reduce([]) { |memo, fruit| memo << fruit << f(fruit) }]
Using length as f() -
2.1.5 :026 > array = ["apples", "bananas", "coconuts", "watermelons"]
=> ["apples", "bananas", "coconuts", "watermelons"]
2.1.5 :027 > Hash[*array.reduce([]) { |memo, fruit| memo << fruit << fruit.length }]
=> {"apples"=>6, "bananas"=>7, "coconuts"=>8, "watermelons"=>11}
2.1.5 :028 >
in addition to the answer of Vlado Cingel (I cannot add a comment yet, so I added an answer).
Inject can also be used in this way: the block has to return the accumulator. Only the assignment in the block returns the value of the assignment, and an error is reported.
array = ["apples", "bananas", "coconuts", "watermelons"]
hash = array.inject({}) { |h,fruit| h[fruit]= f(fruit); h }