Previous   Next

### Binary Search

In the array implementation of sequential search, we can reduce significantly the total search time for a large set of items by using a search procedure based on applying the standard divide-and-conquer paradigm (see ): Divide the set of items into two parts, determine to which of the two parts the search key belongs, then concentrate on that part. A reasonable way to divide the sets of items into parts is to keep the items sorted, then to use indices into the sorted array to delimit the part of the array being worked on. This search technique is called binary search. Program 12.10 is a recursive implementation of this fundamental strategy. Program 2.2 is a non-recursive implementation of the method—no stack is needed because the recursive method in Program 12.10 ends in a recursive call.

Screenshot shows the subfiles examined by binary search when a small table is searched; Screenshot depicts a larger example. Each iteration eliminates slightly more than one-half of the table, so the number of iterations required is small.

##### Screenshot Binary search

Binary search uses only three iterations to find a search key L in this sample file. On the first call, the algorithm compares L to the key in the middle of the file, the G. Since L is larger, the next iteration involves the right half of the file. Then, since L is less than the M in the middle of the right half, the third iteration involves the subfile of size 3 containing H, I, and L. After one more iteration, the subfile size is 1, and the algorithm finds the L.

##### Screenshot Binary search

With binary search, we need only seven iterations to find a record in a file of 200 elements. The subfile sizes follow the sequence 200, 99, 49, 24, 11, 5, 2, 1; each is slightly less than one-half of the previous.

Property 12.5

Binary search never uses more than lg N + 1 comparisons for a search (hit or miss).

See Property 2.3. It is amusing to note that the maximum number of comparisons used for a binary search in a table of size N is precisely the number of bits in the binary representation of N, because the operation of shifting 1 bit to the right converts the binary representation of N into the binary representation of N/2 (see Screenshot).

Keeping the table sorted as we do in insertion sort leads to a running time that is a quadratic function of the number of insert operations, but this cost might be tolerable or even negligible if the number of search operations is huge. In the typical situation where all the items (or even a large number of them) are available before the search begins, we might construct the symbol table by having a constructor that takes an array as its parameter and uses one of the standard sorting methods from Chapters 6 through 10 to sort the table during initialization. After doing so, we could handle updates to the table in various ways. For example, we could maintain order during insert, as in Program 12.8 (see also Exercise 12.25), or we could batch them, sort, and merge (as discussed in Exercise 8.1). Any update could involve an item with a smaller key than those of any item in the table, so every item in the table might have to be moved to make room. This potential for a high cost of updating the table is the biggest liability of using binary search. On the other hand, there are a great many apps where a static table can be presorted and the fast access provided by implementations like Program 12.10 makes binary search the method of choice.

If we need to insert new items dynamically, it seems that we need a linked structure, but a singly linked list does not lead to an efficient implementation, because the efficiency of binary search depends on our ability to get to the middle of any subarray quickly via indexing, and the only way to get to the middle of a singly linked list is to follow links. To combine the efficiency of binary search with the flexibility of linked structures, we need more complicated data structures, which we shall begin examining shortly.

### Binary search (for array-based symbol table)

This implementation of search uses a recursive binary-search procedure. To find whether a given key v is in an ordered array, it first compares v with the element at the middle position. If v is smaller, then it must be in the first half of the array; if v is greater, then it must be in the second half of the array.

The array must be in sorted order. This method could replace search in Program 12.8, which maintains the order dynamically during insertion. We could also include a symbol-table constructor that takes an array as a parameter, then builds a symbol table from items in the parameter array and uses a standard sort routine to prepare it for searching (see Program 12.14).

private ITEM searchR(int l, int r, KEY v) { if (l > r) return null; int m = (l+r)/2; if (equals(v, st[m].key())) return st[m]; if (less(v, st[m].key())) return searchR(l, m-1, v); else return searchR(m+1, r, v); } ITEM search(KEY v) { return searchR(0, N-1, v); }

If duplicate keys are present in the table, then we can extend binary search to support symbol-table operations for counting the number of items with a given key or returning them as a group. Multiple items with keys equal to the search key in the table form a contiguous block in the table (because it is in order), and a successful search in Program 12.10 will end somewhere within this block. If an app requires access to all such items, we can add code to scan both directions from the point where the search terminated and to return two indices delimiting the items with keys equal to the search key. In this case, the running time for the search will be proportional to lg N plus the number of items found. A similar approach solves the more general range-search problem of finding all items with keys falling within a specified interval. We shall consider such extensions to the basic set of symbol-table operations in Part 6.

The sequence of comparisons made by the binary search algorithm is predetermined: The specific sequence used depends on the value of the search key and on the value of N. The comparison structure can be described by a binary-tree structure such as the one illustrated in Screenshot. This tree is similar to the tree that we used in to describe the subfile sizes for mergesort (Screenshot). For binary search, we take one path through the tree; for mergesort, we take all paths through the tree. This tree is static and implicit; in , we shall see algorithms that use a dynamic, explicitly constructed binary-tree structure to guide the search.

##### Screenshot Comparison sequence in binary search

These divide-and-conquer tree diagrams depict the index sequence for the comparisons in binary search. The patterns are dependent on only the initial file size, rather than on the values of the keys in the file. They differ slightly from the trees corresponding to mergesort and similar algorithms (Figures 5.6 and 8.3) because the element at the root is not included in the subtrees. The top diagram shows how a file of 15 elements, indexed from 0 to 14, is searched. We look at the middle element (index 7), then (recursively) consider the left subtree if the element sought is less, and the right subtree if the element sought is greater. Each search corresponds to a path from top to bottom in the tree; for example, a search for an element that falls between the elements 10 and 11 would involve the sequence 7, 11, 9, 10. For file sizes that are not 1 less than a power of 2, the pattern is not quite as regular, as indicated by the bottom diagram, for 12 elements.

One improvement that is possible for binary search is to guess where the search key falls within the current interval of interest with more precision (rather than blindly testing it against the middle element at each step). This tactic mimics the way we look up a name in the telephone directory or a word in a dictionary: If the entry that we are seeking begins with a letter near the beginning of the alphabet, we look near the beginning of the tutorial, but if it begins with a letter near the end of the alphabet, we look near the end of the tutorial. To implement this method, called interpolation search, we modify Program 12.10 to mimic this process. To this end, note that the expression (l + r)/2 that we use in binary search is shorthand for l + 1/2(r - l): We compute the middle of the interval by adding one-half of the size of the interval to the left endpoint. Using interpolation search amounts to replacing 1/2 in this formula by an estimate of where the key might be—specifically (v - kl)/(kr - kl), where kl and kr denote the values of a[l].key() and a[r].key(), respectively. Thus, to implement interpolation search, we replace the statement

m = (l+r)/2;

in binary search by

m = l+(v-a[l].key())*(r-l)/(a[r].key()-a[l].key());

This calculation is based on the assumption that the key values are numerical and evenly distributed.

For files of random keys, it is possible to show that interpolation search uses fewer than lg lg N + 1 comparisons for a search (hit or miss). That proof is quite beyond the scope of this tutorial. This function grows very slowly and can be thought of as a constant for practical purposes: If N is 1 billion, lg lg N < 5. Thus, we can find any item using only a few accesses (on the average)—a substantial improvement over binary search. For keys that are more regularly distributed than random, the performance of interpolation search is even better. Indeed, the limiting case is the key-indexed search method of .

Interpolation search, however, does depend heavily on the assumption that the keys are well distributed over the interval—it can be badly fooled by poorly distributed keys, which do commonly arise in practice. Also, it requires extra computation. For small N, the lg N cost of straight binary search is close enough to lg lg N that the cost of interpolating is not likely to be worthwhile. On the other hand, interpolation search certainly should be considered for large files, for apps where comparisons are particularly expensive, and for external methods where high access costs are involved.

#### Exercises

12.38 Give a nonrecursive implementation of the search method in Program 12.10 (see Program 2.2).

Draw trees that correspond to Screenshot for N = 17 and N = 24.

Find the values of N for which binary search in a symbol table of size N becomes 10, 100, and 1000 times faster than sequential search. Predict the values with analysis and verify them experimentally.

Suppose that insertions into a dynamic symbol table of size N are implemented as in insertion sort, but that we use binary search for search. Assume that searches are 1000 times more frequent than insertions. Estimate the percentage of the total time that is devoted to insertions, for N = 103, 104, 105, and 106.

Develop a symbol-table implementation using binary search and lazy insertion that supports the construct, count, search, insert, and sort operations, using the following strategy. Keep a large sorted array for the main symbol table and an unordered array for recently inserted items. When search is called, sort the recently inserted items (if there are any), merge them into the main table, then use binary search.

Implement a method similar to binary search (Program 12.10) that returns the number of items in the symbol table with keys equal to a given key.

Write a program that, given a value of N, produces a sequence of N macro instructions, indexed from 0 to N-1, of the form compare(l, h), where the ith instruction on the list means "compare the search key with the value at table index i; then report a search hit if equal, do the lth instruction next if less, and do the hth instruction next if greater" (reserve index 0 to indicate search miss). The sequence should have the property that any search will do the same comparisons as would binary search on the same data.

12.47 Develop an expansion of the macro in Exercise 12.46 such that your program produces machine code that can do binary search in a table of size N with as few machine instructions per comparison as possible.

Suppose that a[i] == 10*i for i between 1 and N. How many table positions are examined by interpolation search during the unsuccessful search for 2k - 1?

12.49 Find the values of N for which interpolation search in a symbol table of size N becomes 1, 2, and 10 times faster than binary search, assuming the keys to be random. Predict the values with analysis, and verify them experimentally.

 Previous   Next