I used the top-down psuedo-code from Wikipedia when creating this test code:
public static void main(String args[]) {
int[] nums = new int[] { 17, 5, 3, 7, 6, 3, 11, 2 };
mergeSort(nums);
for (int i = 0; i < nums.length; i++) {
System.out.print(nums[i] + " ,");
}
}
public static void mergeSort(int[] A) {
int[] B = new int[A.length];
System.arraycopy(A, 0, B, 0, A.length);
splitMerge(B, 0, A.length, A); // sort data from B[] into A[]
}
public static void splitMerge(int[] B, int begin, int end, int[] A) {
if (end - begin < 2)
return;
int middle = (end + begin) / 2;
splitMerge(B, begin, middle, A);
splitMerge(B, middle, end, A);
System.out.println("begin: " + begin + " mid: " + ((end - begin)/2) + " end: " + end + " SIZE: " + (end-begin));
merge(B, begin, middle, end, A);
}
public static void merge(int[] B, int begin, int middle, int end, int[] A) {
int i = begin;
int j = middle;
for (int k = begin; k < end; k++) {
if (i < middle && (j >= end || B[i] <= B[j])) {
A[k] = B[i];
i = i + 1;
} else {
A[k] = B[j];
j = j + 1;
}
}
}
This should be more or less exactly the same as the psuedo-code, so why is it not working? Here is the output:
begin: 0 mid: 1 end: 2 SIZE: 2
begin: 2 mid: 1 end: 4 SIZE: 2
begin: 0 mid: 2 end: 4 SIZE: 4
begin: 4 mid: 1 end: 6 SIZE: 2
begin: 6 mid: 1 end: 8 SIZE: 2
begin: 4 mid: 2 end: 8 SIZE: 4
begin: 0 mid: 4 end: 8 SIZE: 8
6 ,3 ,11 ,2 ,17 ,5 ,3 ,7 ,
Your B
array contains a copy of the original array and you never change it, while you change your original A
array.
This means that the merge
method cannot work, since it merges two unsorted arrays (two sub-arrays of B
, which remains unsorted throughout the algorithm).
In order to do the merge
properly, you should copy the elements having indices from begin
to end
from the A
array to the B
array at the start of the merge method.
All you have to do is add the statement
System.arraycopy(A, begin, B, begin, end - begin);
as the first statement of your merge
method.
i.e.:
public static void merge(int[] B, int begin, int middle, int end, int[] A) {
System.arraycopy(A, begin, B, begin, end - begin);
int i = begin;
int j = middle;
for (int k = begin; k < end; k++) {
if (i < middle && (j >= end || B[i] <= B[j])) {
A[k] = B[i];
i = i + 1;
} else {
A[k] = B[j];
j = j + 1;
}
}
}
This will produce the output
2 ,3 ,3 ,5 ,6 ,7 ,11 ,17 ,
for the given input.