• Recognize that coordinate assignments grow nested Arrays
• Recognize single-coordinate access of Arrays returns an Array

Now that you've had a chance to get familiar with the basics of working with nested Arrays or "matrices" (the plural of "matrix"), or "AoAs," let's cover a "funny" case.

Given the following Ruby code:

spice_rack = [
  ["Mace", "Ginger", "Marojam"],          
  ["Paprika", "Fajita Mix", "Coriander"], 
  ["Parsley", "Sage", "Rosemary"]         
]

You're clear now that we can update by saying spice_rack[1][1] = "Extract of Pizza". But what would happen if we said spice_rack[1][100] = "Poodle Dinner". What would spice_rack look like then?

TRY IT OUT IN IRB. We can't emphasize this behavior enough. Programmers, when they reach head-scratchers like this, always ask IRB to teach them. Your learning will not gel as well if you do not take this step. Don't set your learning backwards!

As we hope you discovered, we can add at any coordinate pair. If the element you need add needs the inner Array to "grow" in order to accommodate it, Ruby will "grow" the Array and fill in the "in-between" values with nil.

spice_rack = [
  ["Mace", "Ginger", "Marojam"],             
  ["Paprika", "Fajita Mix", "Coriander"],    
  ["Parsley", "Sage", "Rosemary"]            
]
spice_rack[1][10] = "Cucumber Water"
spice_rack #=> [
  ["Mace", "Ginger", "Marojam"],
  ["Paprika", "Fajita Mix", "Coriander", nil, nil, nil, nil, nil, nil, nil, "Cucumber Water"],
  ["Parsley", "Sage", "Rosemary"]
]

This isn't a new fact, but some students forget that the inner Arrays are still Arrays upon which we can do all the usual stuff we already know how to do on Arrays.

As such, you can use operators like << ("shovel") on an inner Array:

spice_rack = [
  ["Mace", "Ginger", "Marojam"],          
  ["Paprika", "Fajita Mix", "Coriander"], 
  ["Parsley", "Sage", "Rosemary"]         
]

spice_rack[2] << "Saffron" #=> ["Parsley", "Sage", "Rosemary", "Saffron"]

spice_rack #=> [
  ["Mace", "Ginger", "Marojam"],          
  ["Paprika", "Fajita Mix", "Coriander"], 
  ["Parsley", "Sage", "Rosemary", "Saffron"]         
]

Similarly, if you want to replace a whole Array within the containing Array, you can do so by using one coordinate.

spice_rack = [
  ["Mace", "Ginger", "Marojam"],          
  ["Paprika", "Fajita Mix", "Coriander"], 
  ["Parsley", "Sage", "Rosemary"]         
]

# Spice up your life!
spice_rack[0] = ["Posh", "Scary", "Sporty", "Baby", "Ginger"]

spice_rack #=> [
  ["Posh", "Scary", "Sporty", "Baby", "Ginger"],
  ["Paprika", "Fajita Mix", "Coriander"],
  ["Parsley", "Sage", "Rosemary"]
]

Keep in mind, if a matrix starts off with the same number of rows and elements ("is square"), there's nothing wrong with breaking that, if you need. Ruby won't complain.

We've covered two "tangent" topics about working with nested arrays. We've seen these ideas occasionally cause bugs for learners ("Wait, what? How did all those nils get in there?"). Or we've seen learners get stuck because they stop thinking of inner Arrays as the Arrays they already know and love. They get all tangled up in the coordinate syntax and forget what they already know.