I've been trying to figure out how to write this ruby code more eloquently. Does someone have a better solution?
a[:new] = {} if a[:new].nil?
a[:new].merge!( { new_key => new_value } )
is there a way to write this in a more elegant way? I come across this a lot when dealing with nested hashes that need to check whether an key exist and if not, create it.
Write it as below taking the help of Hash#to_h and NilClass#to_h
a[:new] = a[:new].to_h.merge( { new_key => new_value } )
Example :
hsh1[:a] # => nil
hsh1[:a] = hsh1[:a].to_h.merge({1=>2})
hsh1[:a] # => {1=>2}
hsh2 = {:a => {'k' => 2}}
hsh2[:a] # => {"k"=>2}
hsh2[:a] = hsh2[:a].to_h.merge({1=>2})
hsh2 # => {:a=>{"k"=>2, 1=>2}}
Do this at the beginning:
a = Hash.new{|h, k| h[k] = {}}
then, without caring whether a has a key :new or not, do
a[:new].merge!(new_key => new_value)
or
a[:new][new_key] = new_value
Related
Let's consider this as our hash c = {:test => {:foo => true}}
Normally, if we would like to print value of foo, we would access the hash like this c[:test][:foo] but I would like to access it dynamically based on my variable.
Therefore, let's consider the following variable path = [[:test],[:foo]].
How do I access the value true now? I have tried c[path] but it just says nil. What am I missing?
You can make use of dig. You can check the doc of dig here Hash#dig
c = { :test => { :foo => true } }
c[:test][:foo]
#=> true
c.dig(:test, :foo)
#=> true
path = [:test, :foo]
c.dig(*path)
#=> true
You just need to pass the hierarchy
Note: The * before path in c.dig(*path) is reffered as splat operator
Old good recursive Ruby 1.9+ solution:
hash = {:test => {:foo => true}}
path = [[:test],[:foo]]
path.flatten.reduce(hash) { |h, p| h[p] }
#⇒ true
Or, as #Stefan suggested in comments:
path.reduce(hash) { |h, (p)| h[p] }
# or even
path.reduce(hash) { |h, p| h[p.first] }
More defensive:
path.flatten.reduce(hash) { |h, p| h.nil? ? nil : h[p] }
Here is what I am doing now:
def get_counts
products = Product.all
a_count, b_count, c_count = 0, 0, 0
products.collect{ |p| a_count+=1 if p.some_attribute == 'a' }
products.collect{ |p| b_count+=1 if p.some_attribute == 'b' }
products.collect{ |p| c_count+=1 if p.some_attribute == 'c' }
return a_count, b_count, c_count
end
This feels horribly scripty to me. I tried using inject but couldn't get it to work how I wanted. Does anyone have a better way to do this?
To improve on #xdazz's answer
def get_counts
Product.where(some_attribute: ['a','b','c']).
count(group: "some_attribute")
end
This will return a hash in the form:
{'a' => 3, 'b' => 4, 'c' => 5}
def get_counts
return Product.where(:some_attribute => 'a').count, Product.where(:some_attribute => 'b').count, Product.where(:some_attribute => 'c').count
end
If you want only one query, then use group by.
I am using Ruby on Rails 3.2.2 and Ruby 1.9.3. I would like to compare two Hashes (A and B) so to return true if a Hash (A) include all keys/values of the other Hash (B).
For example, given I have
params.inspect
# => { "action"=>"...", "controller"=>"...", "key_param1"=>"value_param1", , "key_param2"=>"value_param2", "key_param3"=>"value_param3", ... }
my_hash1.inspect
# => { "key_param1"=>"value_param1", "key_param2"=>"value_param2" }
my_hash2.inspect
# => { "key_param4"=>"value_param4", "key_param1"=>"value_param1" }
my_hash3.inspect
# => {}
Then I am looking for a method (or something like that) in order to make
params.has_same_keys_and_values_as?(my_hash1)
# => true
params.has_same_keys_and_values_as?(my_hash2)
# => false
params.has_same_keys_and_values_as?(my_hash3)
# => true
Assuming that Hash#keys and Hash#values return values in the same order:
params.values_at(*my_hash.keys) == my_hash.values
I think you can use:
a.slice(*b.keys) == b
where a and b are your hashes. note that slice is a rails method and not ruby.
in plain ruby you can write:
a.keep_if{|k, v| b[k]} == b
class Hash
def >=(b)
eq = true
b.each { |k, v| eq &= !(self.include? k) ? false : ( ( ((self[k]&&v).is_a? Hash) && !((v||self[k]).empty?) ) ? self[k]>=v : true)}
return eq
end
end
params = { "action"=>"...", "controller"=>"...", "key_param1"=>"value_param1", "key_param2"=>"value_param2", "key_param3"=>"value_param3" }
my_hash1 = { "key_param1"=>"value_param1", "key_param2"=>"value_param2" }
my_hash2 = { "key_param4"=>"value_param4", "key_param1"=>"value_param1" }
my_hash3 = {}
p params >= my_hash1 #true
p params >= my_hash2 #false
p params >= my_hash3 #true
It'll work with "deep" hashes as well:
b = {1 => {2 => {} }, 4 => {} }
a = {1 => {2 => {3 => {} }}, 4 => {}, 5 => "123" }
p a >= b #true
p b >= a #false
P.S.
Whether one hash includes another hash
EDIT: This is assuming that the values/keys are not in the same order in both hashes.
You could iterate over each key in hash1 and use has_key? on hash2. Keep in mind this is naive solution that could be slow for large datasets.
Checkout has_key? and has_value? here: http://www.ruby-doc.org/core-1.9.3/Hash.html#method-i-has_key-3F
You could loop as follows:
hash1.each_key { |key|
if hash2.has_key?(key)
do whatever
endif
}
better way, there's an active support method for this, hash.diff, wrap it with .empty? to check if they are the same
{:one => 1}.diff({:one => 1}).empty?
=> true
{:one => 1}.diff({:one => 2}).empty?
=> false
http://as.rubyonrails.org/classes/ActiveSupport/CoreExtensions/Hash/Diff.html
I have an array that contains dates and values. An example of how it might look:
[
{'1/1/2010' => 'aa'},
{'1/1/2010' => 'bb'},
{'1/2/2010' => 'cc'},
{'1/2/2010' => 'dd'},
{'1/3/2010' => 'ee'}
]
Notice that some of the dates repeat. I'm trying to output this in a table format and I only want to show unique dates. So I loop through it with the following code to get my desired output.
prev_date = nil
#reading_schedule.reading_plans.each do |plan|
use_date = nil
if plan.assigned_date != prev_date
use_date = plan.assigned_date
end
prev_date = plan.assigned_date
plan.assigned_date = use_date
end
The resulting table will then look something like this
1/1/2010 aa
bb
1/2/2010 cc
dd
1/3/2010 ee
This work fine but I am new to ruby and was wondering if there was a better way to do this.
Enumerable.group_by is a good starting point:
require 'pp'
asdf = [
{'1/1/2010' => 'aa'},
{'1/1/2010' => 'bb'},
{'1/2/2010' => 'cc'},
{'1/2/2010' => 'dd'},
{'1/3/2010' => 'ee'}
]
pp asdf.group_by { |n| n.keys.first }.map{ |a,b| { a => b.map { |c| c.to_a.last.last } } }
# >> [{"1/1/2010"=>["aa", "bb"]}, {"1/2/2010"=>["cc", "dd"]}, {"1/3/2010"=>["ee"]}]
Which should be a data structure you can bend to your will.
I don't know as though it's better, but you could group the values by date using (e.g.) Enumerable#reduce (requires Ruby >= 1.8.7; before that, you have Enumerable#inject).
arr.reduce({}) { |memo, obj|
obj.each_pair { |key, value|
memo[key] = [] if ! memo.has_key?(key);
memo[key] << value
}
memo
}.sort
=> [["1/1/2010", ["aa", "bb"]], ["1/2/2010", ["cc", "dd"]], ["1/3/2010", ["ee"]]]
You could also use Array#each to similar effect.
This is totally a job for a hash.
Create a hash and use the date as the hashkey and an empty array as the hashvalue.
Then accumulate the values from the original array in the hashvalue array
I am wondering what is the best way to convert a json formatted key value pair to ruby hash with symbol as key:
example:
{ 'user': { 'name': 'foo', 'age': 40, 'location': { 'city' : 'bar', 'state': 'ca' } } }
==>
{ :user=>{ :name => 'foo', :age =>'40', :location=>{ :city => 'bar', :state=>'ca' } } }
Is there a helper method can do this?
using the json gem when parsing the json string you can pass in the symbolize_names option. See here: http://flori.github.com/json/doc/index.html (look under parse)
eg:
>> s ="{\"akey\":\"one\",\"bkey\":\"two\"}"
>> JSON.parse(s,:symbolize_names => true)
=> {:akey=>"one", :bkey=>"two"}
Leventix, thank you for your answer.
The Marshal.load(Marshal.dump(h)) method probably has the most integrity of the various methods because it preserves the original key types recursively.
This is important in case you have a nested hash with a mix of string and symbol keys and you want to preserve that mix upon decode (for instance, this could happen if your hash contains your own custom objects in addition to highly complex/nested third-party objects whose keys you cannot manipulate/convert for whatever reason, like a project time constraint).
E.g.:
h = {
:youtube => {
:search => 'daffy', # nested symbol key
'history' => ['goofy', 'mickey'] # nested string key
}
}
Method 1: JSON.parse - symbolizes all keys recursively => Does not preserve original mix
JSON.parse( h.to_json, {:symbolize_names => true} )
=> { :youtube => { :search=> "daffy", :history => ["goofy", "mickey"] } }
Method 2: ActiveSupport::JSON.decode - symbolizes top-level keys only => Does not preserve original mix
ActiveSupport::JSON.decode( ActiveSupport::JSON.encode(h) ).symbolize_keys
=> { :youtube => { "search" => "daffy", "history" => ["goofy", "mickey"] } }
Method 3: Marshal.load - preserves original string/symbol mix in the nested keys. PERFECT!
Marshal.load( Marshal.dump(h) )
=> { :youtube => { :search => "daffy", "history" => ["goofy", "mickey"] } }
Unless there is a drawback that I'm unaware of, I'd think Method 3 is the way to go.
Cheers
There isn't anything built in to do the trick, but it's not too hard to write the code to do it using the JSON gem. There is a symbolize_keys method built into Rails if you're using that, but that doesn't symbolize keys recursively like you need.
require 'json'
def json_to_sym_hash(json)
json.gsub!('\'', '"')
parsed = JSON.parse(json)
symbolize_keys(parsed)
end
def symbolize_keys(hash)
hash.inject({}){|new_hash, key_value|
key, value = key_value
value = symbolize_keys(value) if value.is_a?(Hash)
new_hash[key.to_sym] = value
new_hash
}
end
As Leventix said, the JSON gem only handles double quoted strings (which is technically correct - JSON should be formatted with double quotes). This bit of code will clean that up before trying to parse it.
Recursive method:
require 'json'
def JSON.parse(source, opts = {})
r = JSON.parser.new(source, opts).parse
r = keys_to_symbol(r) if opts[:symbolize_names]
return r
end
def keys_to_symbol(h)
new_hash = {}
h.each do |k,v|
if v.class == String || v.class == Fixnum || v.class == Float
new_hash[k.to_sym] = v
elsif v.class == Hash
new_hash[k.to_sym] = keys_to_symbol(v)
elsif v.class == Array
new_hash[k.to_sym] = keys_to_symbol_array(v)
else
raise ArgumentError, "Type not supported: #{v.class}"
end
end
return new_hash
end
def keys_to_symbol_array(array)
new_array = []
array.each do |i|
if i.class == Hash
new_array << keys_to_symbol(i)
elsif i.class == Array
new_array << keys_to_symbol_array(i)
else
new_array << i
end
end
return new_array
end
Of course, there is a json gem, but that handles only double quotes.
Another way to handle this is to use YAML serialization/deserialization, which also preserves the format of the key:
YAML.load({test: {'test' => { ':test' => 5}}}.to_yaml)
=> {:test=>{"test"=>{":test"=>5}}}
Benefit of this approach it seems like a format that is better suited for REST services...
The most convenient way is by using the nice_hash gem: https://github.com/MarioRuiz/nice_hash
require 'nice_hash'
my_str = "{ 'user': { 'name': 'foo', 'age': 40, 'location': { 'city' : 'bar', 'state': 'ca' } } }"
# on my_hash will have the json as a hash
my_hash = my_str.json
# or you can filter and get what you want
vals = my_str.json(:age, :city)
# even you can access the keys like this:
puts my_hash._user._location._city
puts my_hash.user.location.city
puts my_hash[:user][:location][:city]
If you think you might need both string and symbol keys:
JSON.parse(json_string).with_indifferent_access