A trie is a search tree that is typically used to store string keys and facilitates fast lookup of values and partial matching. For example, I have a large list of words. Or, say I want to match words that have letters existing in fixed positions but could have any letter in other positions. You can think of a trie as an additional constraint added to a hash table. The position of a letter in a word is given by its distance away from the root. The second letter is at level 2, the third level at level 3, etc.
You might be asking yourself: how do I even go about accessing these words or possibly: how on earth is this structure advantageous? Then you search the Trie using each letter as a key. Tree['H'] brings you to. We actually ran into an edge case for storing data in tries: since multiple pieces of data words can be stored along the same path through the tree, how do we delineate individual words? If a key does not exist at a given level, the word is not in the tree, so we return false.
Note that this algorithm makes adding words to the tree whose paths already exist quite simple. Our Trie is a simple object with nodes and a length. To insert a word into the trie, we split it into individual characters and then convert them all to upper case.
We set the current node equal to the trie root and then check to see if the first letter of the word exists as a key. If it does, we set the current node equal to this next node. We also have a flag that we set to true when we create a new node so that we can increment the length of the trie accordingly.
When we get to the last letter in the word index of word. You might think that deletion is simply deleting the end property from a given node. What makes it lazy? By toggling the end flag, we are potentially leaving extra letters in the tree that no longer belong to any word. No big deal.
