text
stringlengths
1
81
start
float64
0
10.1k
duration
float64
0
24.9
So therein lies our logarithmic running time.
3,903.32
3.14
Therein lies the height of the tree, so long
3,906.46
3.43
as I am good about keeping the tree balanced.
3,909.89
5.88
There's a danger.
3,915.77
1.25
Suppose that I go ahead and start building this tree myself in code
3,917.02
6.42
and I'm a little sloppy about doing that.
3,923.44
2.59
And I go ahead and I insert, for instance, let's say the number 33.
3,926.03
7.88
And it's the first node in my tree, so I'm
3,933.91
2.28
going to put it right up here at the top.
3,936.19
2.69
And now suppose that the next number that just happens
3,938.88
2.45
to get inserted into this tree is 44.
3,941.33
2.225
Well, where does it go?
3,943.555
1.485
Well, it has no children yet, but it is bigger,
3,945.04
2.57
so it should probably go over here.
3,947.61
1.5
So, yeah, I'll draw 44 there.
3,949.11
1.79
Now, suppose that the inputs to this problem
3,950.9
2.36
are such that 55 is inserted next.
3,953.26
1.79
Where does it go?
3,955.05
1.19
All right, 55, it's bigger, so it should go over here.
3,956.24
4.09
And then 66 is inserted next.
3,960.33
2.2
All right, it goes over here-- never mind that.
3,962.53
6.14
So, what's happening to my binary search tree?
3,968.67
3.342
Well, first of all, is it a binary search tree?
3,972.012
1.958
It is because this node is bigger than its left child,
3,973.97
4.51
if any-- there just isn't any-- and it's less than its right child.
3,978.48
3.69
How about here, 44?
3,982.17
1.02
It's bigger than its left child, if any-- because there is none--
3,983.19
2.81
and it's smaller than its right child.
3,986
1.94
The same thing is true for 55, the same thing is true for 66.
3,987.94
2.71
So, this is a binary search tree and yet somehow what does it look like?
3,990.65
5.41
It looks like a linked list, right?
3,996.06
2.27
It's at a weird angle.
3,998.33
1.09
I've been drawing everything horizontally,
3,999.42
1.75
but that's a meaningless artistic detail.
4,001.17
2.12
It devolves potentially into a linked list.
4,003.29
2.67
And so, binary search trees if they are balanced, so to speak,
4,005.96
3.58
if they are built in the right order or built with the right insertion
4,009.54
3.8
algorithm such that they do have this balanced height,
4,013.34
4
this logarithmic height, do afford us the same logarithmic running time
4,017.34
4.49
that the phone book example did and our binary search of an array did.
4,021.83
4.04
But we have to do a little bit more work in order
4,025.87
2.47
to make sure that these trees are balanced.
4,028.34
2.05
And we won't go into detail as to the algorithmics
4,030.39
2.21
of keeping the tree balanced.
4,032.6
2.07
But realize, again, there's going to be this trade-off.
4,034.67
2.44
Yes, you can use a binary search tree or trees more generally to store numbers.
4,037.11
4.03
Yes, they can allow you to achieve that same binary or logarithmic running time
4,041.14
4.78
that we've gotten so used to with arrays,
4,045.92
1.94
but they also give us dynamism such that we
4,047.86
2.34
can keep adding or even removing nodes.
4,050.2
2.39
But, but, but, but it turns out we're going
4,052.59
2.58
to have to think a lot harder about how to keep these things balanced.
4,055.17
4.02
And indeed, in higher level CS courses, courses
4,059.19
2.35
on data structures and algorithms will you
4,061.54
1.75
explore concepts along exactly those lines.
4,063.29
2.96
How would you go about implementing insert and delete into a tree
4,066.25
4.34
so that you do maintain this balance?
4,070.59
2.01
And there is yet more variance on these kinds of trees
4,072.6
2.3
that you'll encounter accordingly.
4,074.9
1.55
But for our purposes, let's consider how you would implement the tree itself
4,076.45
3.59
independent of how you might implement those actual algorithms.
4,080.04
4.27
Let me propose this type of node.
4,084.31
1.74
Again, notice just the very generic term in programming
4,086.05
2.62
where it's usually like a container for one or more other things, and this time
4,088.67
4.436
those things are an integer-- we'll call it n
4,093.106
1.874
but it could be called anything-- and two pointers.
4,094.98
2.87
And instead of next, I'm going to just by convention call them left and right.
4,097.85
4.78
And as before, notice that I do need to declare struct node up
4,102.63
5.13
here or some word up here.
4,107.76
1.769
But by convention I'm just going to do typedef struct node, because C reads
4,109.529
4.431
things top to bottom, left to right.
4,113.96
1.5
So if I want to refer to a node inside of a node,
4,115.46
3.069
I need to have that vocabulary, per this first line, even though later on I
4,118.529
4.381
just want to call this whole darn thing a node.
4,122.91
3.019
And so, that's the distinction.
4,125.929
1.291
This actually has the side effect of creating a data
4,127.22
2.219
type by two different names.
4,129.439
1.166
One is called struct node, and you can literally in your code
4,130.605
2.805
write struct node something, struct node something.
4,133.41
2.44
It just feels unnecessarily verbose, so typedef
4,135.85
3.179
allows you to simplify this as just node, which
4,139.029
2.62
refers to the same structure.
4,141.649
1.371
But this is necessary for this innermost implementation detail.
4,143.02
4.459
So, now that we have the ability with a structure
4,147.479
2.101
to represent this thing, what can we actually do with it?
4,149.58
3.29
Well, here is where recursion from a few weeks ago
4,152.87
3.88
actually gets really compelling.
4,156.75
1.75
When we introduced that sigma example a while ago and talked in the abstract
4,158.5
4.41
about recursion, frankly, it's kind of hard to justify it early on, unless you
4,162.91
4.35
actually have a problem that lends itself to recursion in a way that
4,167.26
3.41
makes sense to use recursion and not just iteration,
4,170.67
2.559
loops-- for loops, while loops, do while, and the like.
4,173.229
2.811
And here we actually have a perfect incarnation of that.
4,176.04
2.739
What does it mean to search a binary search tree?
4,178.779
2.45
Well, suppose I'm searching for a number n
4,181.229
2.781
and I'm being given a pointer to the root of the tree,
4,184.01
2.75
and I'll call it tree.
4,186.76
1.04
So, just like when I was searching a linked list,
4,187.8
1.63
I'm given two things, the number I'm looking for
4,189.43
1.82
and a pointer to the first thing in the data structure-- the first thing
4,191.25
2.779