Nested Collections

Introduction

This lesson is going to build on your knowledge of arrays and hashes. As you have seen, arrays and hashes are great ways to store data. However, sometimes the data is more complex and needs more structure than a basic array or hash offers.

Let’s take a look at how you can use nested arrays and nested hashes to store more complex data.

Lesson overview

This section contains a general overview of topics that you will learn in this lesson.

Describe a nested array and hash.
Explain what data is useful to store in a nested array and hash.
Explain how to access data in a nested array and hash.
Explain why the #dig method is useful.
Explain how to add data to a nested array and hash.
Explain how to delete data in a nested array and hash.
Explain how to create a new nested array that is not mutable.
Explain how to iterate over a nested array and hash.

Nested arrays

Arrays can contain any type of data, including other arrays. An array that contains other arrays is called a nested array, or a multidimensional array.

Nested arrays can be useful to store groups of similar data or positional data. The following nested array of test scores is storing groups of similar data and the teacher mailboxes is storing groups of positional data.

test_scores = [
  [97, 76, 79, 93],
  [79, 84, 76, 79],
  [88, 67, 64, 76],
  [94, 55, 67, 81]
]

teacher_mailboxes = [
  ["Adams", "Baker", "Clark", "Davis"],
  ["Jones", "Lewis", "Lopez", "Moore"],
  ["Perez", "Scott", "Smith", "Young"]
]

Accessing elements

You already know that every element in an array has an index. Accessing a specific element within a nested array is done by calling array[x][y], where x is the index of the nested element and y is the index inside of the nested element.

teacher_mailboxes[0][0]
#=> "Adams"
teacher_mailboxes[1][0]
#=> "Jones"
teacher_mailboxes[2][0]
#=> "Perez"

You can also use negative indices to return elements from the end of an array, starting at [-1].

teacher_mailboxes[0][-1]
#=> "Davis"
teacher_mailboxes[-1][0]
#=> "Perez"
teacher_mailboxes[-1][-2]
#=> "Smith"

If you try to access an index of a nonexistent nested element, it will raise an NoMethodError, because the nil class does not have a [] method. However, just like a regular array, if you try to access a nonexistent index inside of an existing nested element, it will return nil.

teacher_mailboxes[3][0]
#=> NoMethodError
teacher_mailboxes[0][4]
#=> nil

If you want a nil value returned when trying to access an index of a nonexistent nested element, you can use the #dig method. This method can also be used when accessing a nonexistent index inside of an existing nested element.

teacher_mailboxes.dig(3, 0)
#=> nil
teacher_mailboxes.dig(0, 4)
#=> nil

Creating a new nested array

Now that you have seen how to access values inside a nested array, we need to take a step back to look at creating a new nested array. In a previous lesson you were taught to create a new array, by calling the Array.new method with up to 2 optional arguments (initial size and default value), like Array.new(3) or Array.new(3, 7). However, there is one major “gotcha” that is important to point out. According to the Array class documentation, the second optional argument for Array.new (the default value), should only be used with an immutable (unable to be changed) object such as a number, boolean value, or symbol. Using a string, array, hash, or other mutable object may result in confusing behavior because each default value in the array will actually be a reference to the same default value. Therefore, any change to one of the elements will change all of the elements in the array.

To create an immutable array of mutable objects (string, array, hash, etc), you will need to pass the default value for Array.new via a block, using curly braces, instead of the second optional argument. The code in the block gets evaluated for every slot in the array, creating multiple objects to initialize the array with, rather than references to the same object.

To see this for yourself, let’s look at two examples. This first example uses the second optional argument for the default value.

mutable = Array.new(3, Array.new(2))
#=> [[nil, nil], [nil, nil], [nil, nil]]
mutable[0][0] = 1000
#=> 1000
mutable
#=> [[1000, nil], [1000, nil], [1000, nil]]

Changing the value of the first element in the first nested array, causes the first element to change in all three nested arrays! This same behavior will happen with strings, hashes, or any other mutable objects.

Now, let’s take a look at an example that omits the second optional argument and instead passes in the mutable value in a block.

immutable = Array.new(3) { Array.new(2) }
#=> [[nil, nil], [nil, nil], [nil, nil]]
immutable[0][0] = 1000
#=> 1000
immutable
#=> [[1000, nil], [nil, nil], [nil, nil]]

Changing the value of the first element in the first nested array does not cause the value to change in any other nested array.

Adding and removing elements

You can add another element to the end of nested array using the #push method or the shovel operator <<. If you want to add an element to a specific nested array, you will need to specify the index of the nested array.

test_scores << [100, 99, 98, 97]
#=> [[97, 76, 79, 93], [79, 84, 76, 79], [88, 67, 64, 76], [94, 55, 67, 81], [100, 99, 98, 97]]
test_scores[0].push(100)
#=> [97, 76, 79, 93, 100]
test_scores
#=> [[97, 76, 79, 93, 100], [79, 84, 76, 79], [88, 67, 64, 76], [94, 55, 67, 81], [100, 99, 98, 97]]

Using this similar syntax, you can add or remove elements from the entire nested array or from a specific nested element.

test_scores.pop
#=> [100, 99, 98, 97]
test_scores[0].pop
#=> 100
test_scores
#=> [[97, 76, 79, 93], [79, 84, 76, 79], [88, 67, 64, 76], [94, 55, 67, 81]]

Iterating over a nested array

Let’s break down how to iterate over a nested array using the #each_with_index method. You might find it helpful to think of a nested array as having rows and columns. Each row is the nested element and each column is the index of the nested element. When we iterate over the teacher_mailboxes example, each element will be one row.

teacher_mailboxes.each_with_index do |row, row_index|
  puts "Row:#{row_index} = #{row}"
end
#=> Row:0 = ["Adams", "Baker", "Clark", "Davis"]
#=> Row:1 = ["Jones", "Lewis", "Lopez", "Moore"]
#=> Row:2 = ["Perez", "Scott", "Smith", "Young"]
#=> [["Adams", "Baker", "Clark", "Davis"], ["Jones", "Lewis", "Lopez", "Moore"], ["Perez", "Scott", "Smith", "Young"]]

To iterate over the individual elements inside of each row, you will need to nest another enumerable method inside.

teacher_mailboxes.each_with_index do |row, row_index|
  row.each_with_index do |teacher, column_index|
    puts "Row:#{row_index} Column:#{column_index} = #{teacher}"
  end
end
#=> Row:0 Column:0 = Adams
#=> Row:0 Column:1 = Baker
#=> Row:0 Column:2 = Clark
#=> Row:0 Column:3 = Davis
#=> Row:1 Column:0 = Jones
#=> Row:1 Column:1 = Lewis
#=> Row:1 Column:2 = Lopez
#=> Row:1 Column:3 = Moore
#=> Row:2 Column:0 = Perez
#=> Row:2 Column:1 = Scott
#=> Row:2 Column:2 = Smith
#=> Row:2 Column:3 = Young
#=> [["Adams", "Baker", "Clark", "Davis"], ["Jones", "Lewis", "Lopez", "Moore"], ["Perez", "Scott", "Smith", "Young"]]

Although these examples are a bit contrived, it is important to note that if we only need each teacher’s name it would be beneficial to use #flatten before iterating.

teacher_mailboxes.flatten.each do |teacher|
  puts "#{teacher} is amazing!"
end
#=> Adams is amazing!
#=> Baker is amazing!
#=> Clark is amazing!
#=> Davis is amazing!
#=> Jones is amazing!
#=> Lewis is amazing!
#=> Lopez is amazing!
#=> Moore is amazing!
#=> Perez is amazing!
#=> Scott is amazing!
#=> Smith is amazing!
#=> Young is amazing!
#=> ["Adams", "Baker", "Clark", "Davis", "Jones", "Lewis", "Lopez", "Moore", "Perez", "Scott", "Smith", "Young"]

Now let’s take a look at a more complicated example of nesting two predicate enumerables together. Using the above nested array of test scores, let’s determine if any student scored higher than 80 on everything.

test_scores = [[97, 76, 79, 93], [79, 84, 76, 79], [88, 67, 64, 76], [94, 55, 67, 81]]
#=> [[97, 76, 79, 93], [79, 84, 76, 79], [88, 67, 64, 76], [94, 55, 67, 81]]

test_scores.any? do |scores|
  scores.all? { |score| score > 80 }
end
#=> false

This seems pretty straight-forward. It returns false, because none of the nested arrays have scores that are all over 80. What do you think will happen if we switch #any? and #all?? Do you think we will get the same results?

test_scores.all? do |scores|
  scores.any? { |score| score > 80 }
end
#=> true

The results are different, because now it is determining if all of the nested arrays contain any number over 80. This returns true, because each of the nested arrays have at least one number over 80.

Nested hashes

The hashes that you’ve seen so far have single key/value pairs. However, just like arrays, they can be nested, or multidimensional. Nested hashes are a very common way to store complex associated data.

vehicles = {
  alice: {year: 2019, make: "Toyota", model: "Corolla"},
  blake: {year: 2020, make: "Volkswagen", model: "Beetle"},
  caleb: {year: 2020, make: "Honda", model: "Accord"}
}

Accessing data

Accessing a specific element in a nested hash is very similar to a nested array. It is done by calling hash[:x][:y], where :x is the key of the hash and :y is the key of the nested hash.

vehicles[:alice][:year]
#=> 2019
vehicles[:blake][:make]
#=> "Volkswagen"
vehicles[:caleb][:model]
#=> "Accord"

Similar to nested arrays, if you try to access a key in a nonexistent nested hash, it will raise an NoMethodError, therefore you may want to use the #dig method. As expected, if you try to access a nonexistent key in an existing nested hash, it will return nil.

vehicles[:zoe][:year]
#=> NoMethodError
vehicles.dig(:zoe, :year)
#=> nil
vehicles[:alice][:color]
#=> nil
vehicles.dig(:alice, :color)
#=> nil

Adding and removing data

You can add more nested hashes, just like a regular hash. Let’s say Dave just bought a new vehicle and we want to add it to the list.

vehicles[:dave] = {year: 2021, make: "Ford", model: "Escape"}
#=> {:year=>2021, :make=>"Ford", :model=>"Escape"}
vehicles
#=> {:alice=>{:year=>2019, :make=>"Toyota", :model=>"Corolla"}, :blake=>{:year=>2020, :make=>"Volkswagen", :model=>"Beetle"}, :caleb=>{:year=>2020, :make=>"Honda", :model=>"Accord"}, :dave=>{:year=>2021, :make=>"Ford", :model=>"Escape"}}

You can also add an element to one of the nested hashes. Let’s say that Dave really loves his new Escape and thinks we should keep track of the color of the vehicles. To add a new key/value pair to a nested hash, specify the key of the nested hash right before naming the new key.

vehicles[:dave][:color] = "red"
#=> "red"
vehicles
#=> {:alice=>{:year=>2019, :make=>"Toyota", :model=>"Corolla"}, :blake=>{:year=>2020, :make=>"Volkswagen", :model=>"Beetle"}, :caleb=>{:year=>2020, :make=>"Honda", :model=>"Accord"}, :dave=>{:year=>2021, :make=>"Ford", :model=>"Escape", :color=>"red"}}

Deleting one of the nested hashes will be just like a regular hash. Let’s say Blake has decided to sell his Beetle and backpack across Europe. So, let’s delete Blake’s car.

vehicles.delete(:blake)
#=> {:year=>2020, :make=>"Volkswagen", :model=>"Beetle"}
vehicles
#=> {:alice=>{:year=>2019, :make=>"Toyota", :model=>"Corolla"}, :caleb=>{:year=>2020, :make=>"Honda", :model=>"Accord"}, :dave=>{:year=>2021, :make=>"Ford", :model=>"Escape", :color=>"red"}}

To delete one of the key/value pairs inside of a nested hash, you first specify the key of the hash. Let’s say Dave decided that we don’t need to specify the color of his vehicle. Therefore, we specify vehicles[:dave] before we indicate the key of the nested hash to delete.

vehicles[:dave].delete(:color)
#=> "red"
vehicles
#=> {:alice=>{:year=>2019, :make=>"Toyota", :model=>"Corolla"}, :caleb=>{:year=>2020, :make=>"Honda", :model=>"Accord"}, :dave=>{:year=>2021, :make=>"Ford", :model=>"Escape"}}

Methods

There are many helpful methods to use with nested hashes. Once you know what data you need from a nested hash, you might find that browsing through the documentation and experimenting with them in IRB is the best way for you to understand how they work.

Let’s look at an example using the vehicles nested hash. Let’s say that we want to know who owns vehicles that are from 2020 or newer. At first glance in the documentation, it looks like #select would be a great method to use.

vehicles.select { |name, data| data[:year] >= 2020 }
#=> {:caleb=>{:year=>2020, :make=>"Honda", :model=>"Accord"}, :dave=>{:year=>2021, :make=>"Ford", :model=>"Escape"}}

Yes, using #select gives us the information that we need. However, what if we only want the names of the owners and not another nested hash. Let’s keep looking through the documentation to see what else we find. The #collect method sounds very useful for this situation. Let’s collect the names if the year is 2020 or newer.

vehicles.collect { |name, data| name if data[:year] >= 2020 }
#=> [nil, :caleb, :dave]

Using #collect gets us a lot closer to only having the names of the new vehicle owners. If you look at this method in the documentation you will see that #collect and #map have the same functionality. Both of these methods use the return value of each iteration, so when the if statement is false, it will return a nil value.

Nil values can cause problems down the road, so let’s look through the documentation to see if we can find a method to help solve this problem. The #compact method returns an array (or hash) without nil values, so let’s chain it on the end of the block.

vehicles.collect { |name, data| name if data[:year] >= 2020 }.compact
#=> [:caleb, :dave]

Yes, using #collect and #compact returns the data that we want. As you can see, chaining methods can be very useful. However, if we continue exploring the documentation, we will find another method that combines the functionality of these two methods. Ruby version 2.7 added a new enumerable method called #filter_map that sounds very useful for this situation.

vehicles.filter_map { |name, data| name if data[:year] >= 2020 }
#=> [:caleb, :dave]

Amazing! We have found a great solution to returning an array that only contains the names of the owners of vehicles from 2020 or newer! Plus, we got experience using other methods that you will probably use in the future. We have found some really useful methods by exploring the documentation when we have a specific use case in mind.

Assignment

Read more about using a hash with a nested array.
This Stack Overflow answer explains more about using arrays with a nested hash.
Complete the nested collections exercises from the ruby-exercises repo.

Knowledge check

The following questions are an opportunity to reflect on key topics in this lesson. If you can’t answer a question, click on it to review the material, but keep in mind you are not expected to memorize or master this knowledge.