Hashes of Hashes Lab

Learning Goals

1. Recognize vocabulary term: "hash of hashes"
2. Recognize that "Hashes of Hashes" are trees
3. Create a Hash of Hashes
4. Read scalar data from a Hash of Hashes
5. Modify scalar data in a Hash of Hashes

Introduction

Having worked with Arrays of Arrays and Arrays of Hashes, you're going to be pleasantly surprised that most of the syntax for Hashes of Hashes (HoH) works like AoAs, but with Hash keys as the components of a coordinate. Unlike an AoA where coordinates help you find a position in a grid, coordinates in HoH can be stacked deeply so that you can tunnel deeply into a HoH.

Key Image: Relationship Tree

To help you when reading the rest of this lesson, keep in mind that a HoH is like a family tree. Every person in a family tree is a, well, person. Every person in a family tree has two chromosomal parents. And those parents are both person instances which means they, in turn have two chromosomal parents...and on...and on....all the way back to the DAWN OF LIFE.

Thus, we can render any given individual like so:

# Not valid code
{
  name: "",
  mother: # THE SAME TYPE OF THING - something with a name, mother, and father key,
  father: # THE SAME TYPE OF THING - something with a name, mother, and father key
}

Because of the potential to go infinitely deep, we'll keep our HoHs only one or two generations wide. Programmers in social networking companies call things like our person instances "nodes."

From Hash to Array of Hashes

Recall that a Hash is like a table, but in code form. It is a way for your program to store pairs of data as elements in a collection.

We call names in a Hash that we use to point to data, keys. By providing a Hash name and a key, we can uniquely identify a value.

Recognize Vocabulary Term: "Hash of Hashes"

HoH's property of being made up of nodes that have the same structure of each other makes them very popular in programming interviews. Much of the vocabulary used to talk of these "node collections" is described by a mathematical discipline called "graph theory." We're not going to discuss that vocabulary in this lesson. But if you hear "leaf," "node," or "edge," you're starting to approach "graph theory."

In the age of social networking being able to build "graphs" of "nodes" has become very important. We'll make a mini-social network in our example below.

Create a HoH

Let's make a simple family tree:

# Simple Hashes
mother_hash = {
  name: "Debbie Reynolds",
  mother: {},
  father: {}
}

father_hash = {
  name: "Eddie Fisher",
  mother: {},
  father: {}
}

# Princess Leia (among other roles)
child = {
  name: "Carrie Fisher",
  mother: mother_hash,
  father: father_hash
} # => {:name=>"Carrie Fisher", :mother=>{:name=>"Debbie Reynolds", :mother=>{}, :father=>{}}, :father=>{:name=>"Eddie Fisher", :mother=>{}, :father=>{}}}

If we don't need to use mother_hash or father_hash, then we don't need to assign them before nesting them in child. We can write a nested HoH using literal formatting:

# AoH in one go
child = {:name=>"Carrie Fisher", :mother=>{:name=>"Debbie Reynolds", :mother=>{}, :father=>{}}, :father=>{:name=>"Eddie Fisher", :mother=>{}, :father=>{}}}

# Same as above, but using whitespace to make it easier for humans to read

child = {
  :name=>"Carrie Fisher",
  :mother=>{
    :name=>"Debbie Reynolds",
    :mother=>{},
    :father=>{}
  },
  :father=>{
    :name=>"Eddie Fisher",
    :mother=>{},
    :father=>{}
  }
}

Since Ruby doesn't care about whitespace, we'd advise you to write a HoH with lots of indentation.

Take a moment to reflect. Can you "see" the tree's nodes? Each entity ("Eddie Fisher", "Debbie Reynolds" and "Carrie Fisher") are the same thing (a person) with another "generation" of Hash data accessible via standard keys. This means that we can "tunnel" through the infinite regress of HoHs to get data.

We've kept the regress very shallow here (2 generations), but you can imagine a longer family tree could be quite deep indeed!

Read Scalar Data From a Nested Array

One of the defining features of HoHs is that we use stacked keys to get data out of the complex tree. To read an element from a HoH simply "stack" keys:

child = {
  :name=>"Carrie Fisher",
  :mother=>{
    :name=>"Debbie Reynolds",
    :mother=>{
      :name=> "Maxene Reynolds"
    },
    :father=>{}
  },
  :father=>{
    :name=>"Eddie Fisher",
    :mother=>{},
    :father=>{}
  }
}

child[:name] #=> "Carrie Fisher"
child[:mother][:name] #=> "Debbie Reynolds"
child[:mother][:mother][:name] #=> "Maxene Reynolds"

Through this "stacking" across "generations" we can see that Carrie Fisher's grandmother's name was "Maxene."

PRO TIP: When building HoHs, it's wise to ensure you have all the data uniform. We should ensure that we have the mother / father data for "Eddie Fisher". To show "end of our knowledge" we've used an empty Hash.

Modify Scalar Data In an Hash of Hashes

The same technique of using coordinates applies both to reading and modification. If you can "target" an element you can read it or update it.

child = {
  :name=>"Carrie Fisher",
  :mother=>{
    :name=>"Debbie Reynolds",
    :mother=>{
      :name=> "Maxene Reynolds"
    },
    :father=>{}
  },
  :father=>{
    :name=>"Eddie Fisher",
    :mother=>{},
    :father=>{}
  }
}

child[:mother][:father] = "Raymond Reynolds"
child #=> {:name=>"Carrie Fisher", :mother=>{:name=>"Debbie Reynolds", :mother=>{:name=>"Maxene Reynolds"}, :father=>"Raymond Reynolds"}, :father=>{:name=>"Eddie Fisher", :mother=>{}, :father=>{}}}
child[:mother][:father] #=> "Raymond Reynolds"

You can also replace inner Hashes with new Hashes:

child = {
  :name=>"Carrie Fisher",
  :mother=>{
    :name=>"Debbie Reynolds",
    :mother=>{
      :name=> "Maxene Reynolds"
    },
    :father=>{
      :name=> "Raymond Reynolds"
    }
  },
  :father=>{
    :name=>"Eddie Fisher",
    :mother=>{},
    :father=>{}
  }
}

child[:father][:mother] = {name: "Gitte Winokur", mother: {}, father: {}}
child #=> {:name=>"Carrie Fisher", :mother=>{:name=>"Debbie Reynolds", :mother=>{:name=>"Maxene Reynolds"}, :father=>{:name=>"Raymond Reynolds"}}, :father=>{:name=>"Eddie Fisher", :mother=>{:name=>"Gitte Winokur", :mother=>{}, :father=>{}}, :father=>{}}}
child[:father][:mother] #=> {name: "Gitte Winokur", mother: {}, father: {}}

While AoAs or AoHs are frequently described using ONLY literal syntax, with a HoH, it might be more readable to define the nodes in single-instance variables and then use the short variables to build the HoH.

HoH in the Wild

As a closing point, it's not a rule that all "nodes" in a HoH are all the same. It's often true, but sometimes not. In the lab, we're going to provide a simple HoH which breaks the tendency of all nodes to be alike.

Instructions

For this lab, your task is to build a nested HoH of the classification of biological life on Earth based on these commented-out snippets assigned to PORTION_1-PORTION_4. Return that HoH as the return value for the method "naming_system."

Take a look at lib/hoh.rb to get started. The tests will guide your construction.

Conclusion

Congratulations, you've learned to use your third nested data structure: the Hash of Hashes. You've seen that you can build them by building Hashes filled with variables that point to Hashes in an infinite regress. You'll start off using HoH very rarely, but, in time, thanks to our networked and social age, you're going to become friends with them.