Bit operations Cracking the Coding Interview

Question

I've been trying to work through the first problem in CTCI, which involves bit manipulation, and I just can't figure out how the author accurately made the mask in his final solution. Could someone explain the calculations for "int left", "int right", and "int mask"? It would be great to see what these lines are calculating specifically for the example he provided.

The Question is: You are given two 32-bit numbers, N and M, and two bit positions, i and j. Write a method to set all bits between i and j in N equal to M (e.g., M becomes a substring of N located at i and starting at j). EXAMPLE: Input: N = 10000000000, M = 10101, i = 2, j = 6 Output: N = 10001010100

public static int updateBits(int n, int m, int i, int j) {
int max = ~0; /* All 1’s */

// 1’s through position j, then 0’s
int left = max - ((1 << j) - 1);

// 1’s after position i
int right = ((1 << i) - 1);

// 1’s, with 0s between i and j
int mask = left | right;

// Clear i through j, then put m in there
return (n & mask) | (m << i);
}

Answer 1

explain the calculations for "int left", "int right", and "int mask"

// 1’s through position j, then 0’s
int left = max - ((1 << j) - 1);

• (1 << j) has just the bit at position j set to 1
• ((1 << j) - 1) (subtracting 1) yields all j bits below (to the right of) position j set to 1
• max - ((1 << j) - 1) (subtracting from all 1) yields the bitwise complement of the above, i. e. all bits above (to the left of) and including position j set to 1, and the j bits below set to 0

e. g.

j          1<<j            (1<<j)-1       ~0-((1<<j)-1)
-------------------------------------------------------
0      000000000001      000000000000      111111111111
1      000000000010      000000000001      111111111110
2      000000000100      000000000011      111111111100
3      000000001000      000000000111      111111111000
4      000000010000      000000001111      111111110000
5      000000100000      000000011111      111111100000
6      000001000000      000000111111      111111000000
...

// 1’s after position i
int right = ((1 << i) - 1);

• (1 << i) has just the bit at position i set to 1
• ((1 << i) - 1) (subtracting 1) yields all i bits below (to the right of) position i set to 1

e. g.

i          1<<i            (1<<i)-1
-------------------------------------
0      000000000001      000000000000
1      000000000010      000000000001
2      000000000100      000000000011
...

// 1’s, with 0s between i and j
int mask = left | right;

i = 2, j = 6:

left      111111000000
right     000000000011
|         ------------
mask      111111000011

// Clear i through j, then put m in there
return (n & mask) | (m << i);

• (n & mask), unlike commented, clears bits i through j-1 only (see mask above)
• (m << i) shifts the value to be set to the desired bit position
• (n & mask) | (m << i) ORs the shifted value bits to the masked n

with your example:

n       010000000000
mask    111111000011
&       ------------
n&mask  010000000000
m<<i    000001010100
|       ------------
        010001010100

Now, although this example values yield a correct result, we can see that the implementation of updateBits() is actually wrong, because the left part of the mask needs 1 bits only to the left of and not including position j, since bit position j belongs to the substring that is to be masked out. The wrongness shows e. g. with the values n = 11111111111, m = 00000, i = 2, j = 6:

n       011111111111
mask    111111000011
&       ------------
n&mask  011111000011
m<<i    000000000000
|       ------------
        011111000011

The value m has been put into bit positions i to j-1 only.

The error can be corrected by changing

int left = max - ((1 << j) - 1);

to

int left = max - ((1 << j+1) - 1);