Lecture 15 Notes

Binary Search Trees

15-122: Principles of Imperative Computation (Fall 2015)

Frank Pfenning, André Platzer, Rob Simmons

1 Introduction
In this lecture, we will continue considering ways to implement the set
(or associative array) interface. This time, we will implement this inter-
face with binary search trees. We will eventually be able to achieve O(log n)
worst-case asymptotic complexity for insert and lookup. This also extends
to delete, although we won’t discuss that operation in lecture.

2 Ordered Associative Arrays

Hashtables are associative arrays that organize the data in an array at an
index that is determined from the key using a hash function. If the hash
function is good, this means that the element will be placed at a reasonably
random position spread out across the whole array. If it is bad, linear search
is needed to locate the element.
There are many alternative ways of implementing associative arrays.
For example, we could have stored the elements in an array, sorted by
key. Then lookup by binary search would have been O(log n), but inser-
tion would be O(n), because it takes O(log n) steps to find the right place,
but then O(n) steps to make room for that new element by shifting all big-
ger elements over. (We would also need to grow the array as in unbounded
arrays to make sure it does not run out of capacity.) Arrays are not flexible
enough for fast insertion, but the data structure that we will be devising in
this lecture will be.

L ECTURE N OTES
Binary Search Trees L15.2

3 Abstract Binary Search

What are the operations that we needed to be able to perform binary search?
We needed a way of comparing the key we were looking for with the key of
a given element in our data structure. Depending on the result of that com-
parison, binary search returns the position of that element if they were the
same, advances to the left if what we are looking for is smaller, or advances
to the right if what we are looking for is bigger. For binary search to work
with the complexity O(log n), it was important that binary search advances
to the left or right many steps at once, not just by one element. Indeed, if we
would follow the abstract binary search principle starting from the middle
of the array but advancing only by one index in the array, we would obtain
linear search, which has complexity O(n), not O(log n).
Thus, binary search needs a way of comparing keys and a way of ad-
vancing through the elements of the data structure very quickly, either to
the left (towards elements with smaller keys) or to the right (towards big-
ger ones). In the array-based binary search we’ve studied, each iteration
calculates a midpoint
int mid = lower + (upper - lower) / 2;
and a new bound for the next iteration is (if the key we’re searching for is
smaller than the element at mid)
upper = mid;
or (if the key is larger)
lower = mid + 1;
So we know that the next value mid0 will be either (lower + mid) / 2 or
((mid + 1) + upper) / 2 (ignoring the possibility of overflow).
This pattern continues, and given any sorted array, we can enumerate
all possible binary searches:
Binary  search  always  looks  here  first…  
…and  looks  at  one  of  these  two  indices  second…  

0   1   2   3   4   5   6   7   8   9   10   11   12   13   14   15   16   17   18  

-­‐9   -­‐6   -­‐1   0   1   5   7   8   9   11   15   16   19   20   29   42   55   88   90  

…and  looks  at  one  of  these  four  indices  third…  

This pattern means that constant-time access to an array element at an

arbitrary index isn’t necessary for doing binary search! To do binary search

L ECTURE N OTES
Binary Search Trees L15.3

on the array above, all we need is constant time access from array index 9
(containing 11) to array indices 4 and 14 (containing 1 and 29, respectively),
constant time access from array index 4 to array indices 2 and 7, and so on.
At each point in binary search, we know that our search will proceed in one
of at most two ways, so we will explicitly represent those choices with an
pointer structure, giving us the structure of a binary tree. The tree structure
that we got from running binary search on this array. . .
0   1   2   3   4   5   6   7   8   9   10   11   12   13   14   15   16   17   18  

-­‐9   -­‐6   -­‐1   0   1   5   7   8   9   11   15   16   19   20   29   42   55   88   90  

. . . corresponds to this binary tree:

11  

29  
1  
19   88  
-­‐1   8  
16   55  
-­‐6   7   20   90  
0   9  
15   42  
-­‐9   5  

4 Representing Binary Trees With Pointers

To represent a binary tree using pointers, we use a struct with two pointers:
one to the left child and one to the right child. If there is no child, the pointer
is NULL. A leaf of the tree is a node with two NULL pointers.

typedef struct tree_node tree;

struct tree_node {
elem data;
tree* left;
tree* right;
};

Rather than the fully generic data implementation that we used for hash
tables, we’ll assume for the sake of simplicity that the client is provid-
ing us with a type of elem that is known to be a pointer, and a single
elem_compare.

/* Client-side interface */
typedef ______* elem;

L ECTURE N OTES
Binary Search Trees L15.4

int elem_compare(elem k1, elem k2)

/*@ensures -1 <= \result && \result <= 1; @*/ ;

The elem_compare function provided by the client is different from the

equivalence function we used for hash tables. For binary search trees, we
need to compare keys k1 and k2 and determine if k1 < k2 , k1 = k2 , or
k1 > k2 . A common approach to this is for a comparison function to re-
turn an integer r, where r < 0 means k1 < k2 , r = 0 means k1 = k2 , and
r > 0 means k1 > k2 . Our contract captures that we expect elem_compare
to return no values other than -1, 0, and 1.
Trees are the second recursive data structure we’ve seen: a tree node has
two fields that contain pointers to tree nodes. Thus far we’ve only seen
recursive data structures as linked lists, either chains in a hash table or list
segments in a stack or a queue.
Let’s remember how we picture list segments. Any list segment is ref-
ered to by two pointers: start and end, and there are two possibilities for
how this list can be constructed, both of which require start to be non-
NULL.

bool is_segment(list* start, list* end) {

if (start == NULL) return false;
if (start == end) return true;
return is_segment(start->next, end);
}

We can represent these choices graphically by using a picture like to

represent an arbitrary segment. Then we know every segment has one or
two forms:

start   start   data   next  

start   data   next  

end   end   end   x  

IS   OR  

We’ll create a similar picture for trees: the tree containing no elements
is NULL, and a non-empty tree is a struct with three fields: the data and the
left and right pointers, which are themselves trees.

data   le)   right  

x  
root   root   root  
IS   OR  

L ECTURE N OTES
Binary Search Trees L15.5

Rather than drawing out the list_node struct with its three fields explicitly,
we’ll usually use a more graph-like way of presenting trees:

x  
root   root   root  
IS   OR  

This recursive definition can be directly encoded into a very simple data
structure invariant is_tree. It checks very little: just that all the data fields
are non-NULL, as the client interface requires. If it terminates, it also en-
sures that there are no cycles; a cycle would cause nontermination, just as
it would with is_segment.
bool is_tree(tree* root) {
if (root == NULL) return true;
return root->data != NULL
&& is_tree(root->left) && is_tree(root->right);
}

4.1 The Ordering Invariant

Binary search was only correct for arrays if the array was sorted. Only then
do we know that it is okay not to look at the upper half of the array if the
element we are looking for is smaller than the middle element, because, in
a sorted array, it can then only occur in the lower half, if at all. For binary
search to work correctly on binary search trees, we, thus, need to maintain
a corresponding data structure invariant: all elements to the right of a node
have keys that are bigger than the key of that node. And all the nodes to the
left of that node have smaller keys than the key at that node. This ordering
invariant is a core idea of binary search trees; it’s what makes a binary tree
into a binary search tree.
Ordering Invariant. At any node with key k in a binary search
tree, all keys of the elements in the left subtree are strictly less
than k, while all keys of the elements in the right subtree are
strictly greater than k.
This invariant implies that no key occurs more than once in a tree, and we
have to make sure our insertion function maintains this invariant.
We won’t write code for checking the ordering invariant just yet, as
that turns out to be suprisingly difficult. We’ll first discuss the lookup and
insertion functions for binary search trees.

L ECTURE N OTES
Binary Search Trees L15.6

5 Searching for a Key

The ordering invariant lets us find an element e in a binary search tree the
same way we found an element with binary search, just on the more ab-
stract tree data structure. Here is a recursive algorithm for search, starting
at the root of the tree:

1. If the tree is empty, stop.

2. Compare the key k of the current node to e. Stop if equal.

3. If e is smaller than k, proceed to the left child.

4. If e is larger than k, proceed to the right child.

The implementation of this search captures the informal description above.

Recall that elem_compare(x,y) returns −1 if x < y, 0 if x = y, and 1 if
x > y.
elem tree_lookup(tree* T, elem x)
//@requires is_tree(T);
{
if (T == NULL) return NULL;
int cmp = elem_compare(x, T->data);
if (cmp == 0) {
return T->data;
} else if (cmp < 0) {
return tree_lookup(T->left, x);
} else {
//@assert cmp > 0;
return tree_lookup(T->right, x);
}
}
We chose here a recursive implementation, following the structure of a tree,
but in practice an iterative version may also be a reasonable alternative (see
Exercise 1).

6 Complexity
If our binary search tree were perfectly balanced, that is, had the same num-
ber of nodes on the left as on the right for every subtree, then the ordering

L ECTURE N OTES
Binary Search Trees L15.7

invariant would ensure that search for an element with a given key has
asymptotic complexity O(log n), where n is the number of elements in the
tree. Every time we compare the element x with the root of a perfectly
balanced tree, we either stop or throw out half the elements in of the tree.
In general we can say that the cost of lookup is O(h), where h is the
height of the tree. We will define height to be the maximum number of
nodes that can be reached by any sequence of pointers starting at the root.
An empty tree has height 0, and a tree with two children has the maximum
height of either child, plus 1.

7 The Interface
Before we talk about insertion into a binary search tree, we should specify
the interface and discuss how we will implement it. Remember that we’re
assuming a single client definition of elem and a single client definition
of elem_compare, rather than the fully generic version using void pointers
and function pointers.

/* Library interface */
typedef ______* bst_t;

bst_t bst_new()
/*@ensures \result != NULL; @*; ;

void bst_insert(bst_t B, elem x)

/*@requires B != NULL && x != NULL; @*/ ;

elem bst_lookup(bst_t B, elem x)

/*@requires B != NULL && x != NULL; @*/ ;

We can’t define bst_t to be tree*, for two reasons. One reason is that a
new tree should be empty, but an empty tree is represented by the pointer
NULL, which would violate the bst_new postcondition. More fundamen-
tally, if NULL was the representation of an empty tree, there would be no
way to imperatively insert additional elements in the tree.
The usual solution here is one we have already used for stacks, queues,
and hash tables: we have a header which in this case just consists of a pointer
to the root of the tree. We often keep other information associated with the
data structure in these headers, such as the size.

L ECTURE N OTES
Binary Search Trees L15.8

typedef struct bst_header bst;

struct bst_header {
tree* root;
};

bool is_bst(bst* B) {
return B != NULL && is_tree(B->root);
}

Lookup in a binary search tree then just calls the recursive function we’ve
already defined:

elem bst_lookup(bst* B, elem x)

//@requires is_bst(B) && x != NULL;
{
return tree_lookup(B->root, x);
}

The relationship between both is_bst and is_tree and between bst_lookup
and tree_lookup is a common one. The non-recursive function is_bst
is given the non-recursive struct bst_header, and then calls the recursive
helper function is_tree on the recursive structure of tree nodes.

8 Inserting an Element
With the header structure, it is straightforward to implement bst_insert.
We just proceed as if we are looking for the given element. If we find a node
with an equivalent element, we just overwrite its data field. Otherwise, we
insert the new key in the place where it would have been, had it been there
in the first place. This last clause, however, creates a small difficulty. When
we hit a null pointer (which indicates the key was not already in the tree),
we cannot replace what it points to (it doesn’t point to anything!). Instead,
we return the new tree so that the parent can modify itself.

L ECTURE N OTES
Binary Search Trees L15.9

tree* tree_insert(tree* T, elem x)

//@requires is_tree(T) && x != NULL;
//@ensures is_tree(\result);
{
if (T == NULL) {
/* create new node and return it */
T = alloc(struct tree_node);
T->data = x;
T->left = NULL; // Not required (initialized to NULL)
T->right = NULL; // Not required (initialized to NULL)
return R;
} else {
int cmp = elem_compare(x, T->data);
if (cmp == 0) {
T->data = x;
} else if (cmp < 0) {
T->left = tree_insert(T->left, x);
} else {
//@assert cmp > 0;
T->right = tree_insert(T->right, x);
}
}

return T;
}

The returned subtree will also be stored as the new root:

void bst_insert(bst* B, elem x)

//@requires is_bst(B)
//@requires x != NULL;
//@ensures is_bst(B);
{
B->root = tree_insert(B->root, x);
}

L ECTURE N OTES
Binary Search Trees L15.10

9 Checking the Ordering Invariant

When we analyze the structure of the recursive functions implementing
search and insert, we are tempted to try defining a simple, but wrong! or-
dering invariant for binary trees as follows: tree T is ordered whenever

1. T is empty, or

2. T has key k at the root, TL as left subtree and TR as right subtree, and

• TL is empty, or TL ’s key is less than k and TL is ordered; and

• TR is empty, or TR ’s key is greater than k and TR is ordered.

This would yield the following code:

/* THIS CODE IS BUGGY */

bool is_ordered(tree* T) {
if (T == NULL) return true; /* an empty tree is a BST */
elem k = T->data;
return (T->left == NULL
|| (elem_compare(T->left->data), k) < 0
&& is_ordered(T->left)))
&& (T->right == NULL
|| (elem_compare(k, T->right->data)) < 0
&& is_ordered(T->right)));
}

While this should always be true for a binary search tree, it is far weaker
than the ordering invariant stated at the beginning of lecture. Before read-
ing on, you should check your understanding of that invariant to exhibit a
tree that would satisfy the above, but violate the ordering invariant.

L ECTURE N OTES
Binary Search Trees L15.11

There is actually more than one problem with this. The most glaring
one is that following tree would pass this test:

7


5
 11


1
 9


Even though, locally, the key of the left node is always smaller and on the
right is always bigger, the node with key 9 is in the wrong place and we
would not find it with our search algorithm since we would look in the
right subtree of the root.
An alternative way of thinking about the invariant is as follows. As-
sume we are at a node with key k.

1. If we go to the left subtree, we establish an upper bound on the keys in

the subtree: they must all be smaller than k.

2. If we go to the right subtree, we establish a lower bound on the keys in

the subtree: they must all be larger than k.

The general idea then is to traverse the tree recursively, and pass down
an interval with lower and upper bounds for all the keys in the tree. The
following diagram illustrates this idea. We start at the root with an unre-
stricted interval, allowing any key, which is written as (−∞, +∞). As usual
in mathematics we write intervals as (x, z) = {y | x < y and y < z}. At
the leaves we write the interval for the subtree. For example, if there were
a left subtree of the node with key 7, all of its keys would have to be in the

L ECTURE N OTES
Binary Search Trees L15.12

interval (5, 7).

(‐∞,
+∞)

9

(‐∞,
9)


5

(5,
9)
 (9,
+∞)


(‐∞,
5)

7


(5,
7)
 (7,
9)


The only difficulty in implementing this idea is the unbounded inter-

vals, written above as −∞ and +∞. Here is one possibility: we pass not
just the key value, but the particular element from which we can extract the
key that bounds the tree. Since elem must be a pointer type, this allows us
to pass NULL in case there is no lower or upper bound.
bool is_ordered(tree* T, elem lower, elem upper) {
if (T == NULL) return true;
return T->data != NULL
&& (lower == NULL || elem_compare(lower, T->data) < 0)
&& (upper == NULL || elem_compare(T->data, upper) < 0)
&& is_ordered(T->left, lower, T->data)
&& is_ordered(T->right, T->data, upper);
}
This checks all the properies that our earlier is_tree checked, so we
can just implement is_tree in terms of is_ordered:

bool is_tree(tree* T) {
return is_ordered(T, NULL, NULL);
}

bool is_bst(bst B) {
return B != NULL && is_tree(B->root);
}

A word of caution: the is_ordered(T, NULL, NULL) pre- and post-

condition of the function tree_insert is actually not strong enough to

L ECTURE N OTES
Binary Search Trees L15.13

prove the correctness of the recursive function. A similar remark applies

to tree_lookup. This is because of the missing information of the bounds.
We will return to this issue later in the course.

10 The Shape of Binary Search Trees

We have already mentioned that balanced binary search trees have good
properties, such as logarithmic time for insertion and search. The question
is if binary search trees will be balanced. This depends on the order of
insertion. Consider the insertion of numbers 1, 2, 3, and 4.
If we insert them in increasing order we obtain the following trees in
sequence.
1
 1
 1
 1


2
 2
 2


3
 3


4


Similarly, if we insert them in decreasing order we get a straight line along,

always going to the left. If we instead insert in the order 3, 1, 4, 2, we obtain
the following sequence of binary search trees:

3
 3
 3
 3


1
 1
 4
 1
 4


2


Clearly, the last tree is much more balanced. In the extreme, if we insert
elements with their keys in order, or reverse order, the tree will be linear,
and search time will be O(n) for n items.
These observations mean that it is extremely important to pay attention
to the balance of the tree. We will discuss ways to keep binary search trees
balanced in a later lecture.

L ECTURE N OTES
Binary Search Trees L15.14

Exercises
Exercise 1 Rewrite tree_lookup to be iterative rather than recursive.

Exercise 2 Rewrite tree_insert to be iterative rather than recursive. [Hint:

The difficulty will be to update the pointers in the parents when we replace a node
that is null. For that purpose we can keep a “trailing” pointer which should be the
parent of the note currently under consideration.]

Exercise 3 The binary search tree interface only expected a single function for key
comparison to be provided by the client:

int elem_compare(elem k1, elem k2);

An alternative design would have been to, instead, expect the client to provide a
set of elem comparison functions, one for each outcome:

bool elem_equal(elem k1, elem k2);

bool elem_greater(elem k1, elem k2);
bool elem_less(elem k1, elem k2);

What are the advantages and disadvantages of such a design?

L ECTURE N OTES