array.sort源码解释

* fast: it is guaranteed to run in n log(n) time andruns substantially faster on nearly sorted lists. empirical tests showed it tobe as fast as a highly optimized quicksort. a quicksort is generally consideredto be faster than a merge sort but isn't stable and doesn't guarantee n log(n)performance.


* stable: it doesn't reorder equal elements. this isimportant if you sort the same list repeatedly on different attributes. if auser of a mail program sorts the inbox by mailing date and then sorts it bysender, the user naturally expects that the now-contiguous list of messagesfrom a given sender will (still) be sorted by mailing date. this is guaranteedonly if the second sort was stable.


也就是说，优化的归并排序既快速（nlog(n)）又稳定。


对于对象的排序，稳定性很重要。比如成绩单,一开始可能是按人员的学号顺序排好了的,现在让我们用成绩排,那么你应该保证,本来张三在李四前面,即使他们成绩相同,张三不能跑到李四的后面去。


而快速排序是不稳定的，而且最坏情况下的时间复杂度是o(n^2)。


另外，对象数组中保存的只是对象的引用，这样多次移位并不会造成额外的开销，但是，对象数组对比较次数一般比较敏感，有可能对象的比较比单纯数的比较开销大很多。归并排序在这方面比快速排序做得更好，这也是选择它作为对象排序的一个重要原因之一。


排序优化：实现中快排和归并都采用递归方式，而在递归的底层，也就是待排序的数组长度小于7时, 直接使用冒泡排序，而不再递归下去。


分析：长度为6的数组冒泡排序总比较次数最多也就1+2+3+4+5+6=21次，最好情况下只有6次比较。而快排或归并涉及到递归调用等的开销，其时间效率在n较小时劣势就凸显了，因此这里采用了冒泡排序，这也是对快速排序极重要的优化。


/*快速排序*/
private static void sort1(int x[], int off, int len) {
// insertion sort on smallest arrays
if (len < 7) {
for (int i=off; i<len+off; i++)
for (int j=i; j>off && x[j-1]>x[j]; j--)
swap(x, j, j-1);
return;
}
// choose a partition element, v
int m = off + (len >> 1); // small arrays, middle element
if (len > 7) {
int l = off;
int n = off + len - 1;
if (len > 40) { // big arrays, pseudomedian of 9
int s = len/8;
l = med3(x, l, l+s, l+2*s);//取后三个参数中的中间值
m = med3(x, m-s, m, m+s);
n = med3(x, n-2*s, n-s, n);
}
m = med3(x, l, m, n); // mid-size, med of 3
}
int v = x[m];
// establish invariant: v* (<v)* (>v)* v*
int a = off, b = a, c = off + len - 1, d = c;
while(true) {
while (b <= c && x[b] <= v) {
if (x[b] == v)
swap(x, a++, b);
b++;
}
while (c >= b && x[c] >= v) {
if (x[c] == v)
swap(x, c, d--);
c--;
}
if (b > c)
break;
swap(x, b++, c--);
}
// swap partition elements back to middle
int s, n = off + len;
s = math.min(a-off, b-a ); vecswap(x, off, b-s, s);
s = math.min(d-c, n-d-1); vecswap(x, b, n-s, s);
// recursively sort non-partition-elements
if ((s = b-a) > 1)
sort1(x, off, s);
if ((s = d-c) > 1)
sort1(x, n-s, s);
}
/*归并排序*/
private static void mergesort(object[] src,
object[] dest,
int low,
int high,
int off) {
int length = high - low;
// insertion sort on smallest arrays
if (length < insertionsort_threshold) {
for (int i=low; i<high; i++)
for (int j=i; j>low &&
((comparable) dest[j-1]).compareto(dest[j])>0; j--)
swap(dest, j, j-1);
return;
}
// recursively sort halves of dest into src
int destlow = low;
int desthigh = high;
low += off;
high += off;
int mid = (low + high) >>> 1;
mergesort(dest, src, low, mid, -off);
mergesort(dest, src, mid, high, -off);
// if list is already sorted, just copy from src to dest. this is an
// optimization that results in faster sorts for nearly ordered lists.
if (((comparable)src[mid-1]).compareto(src[mid]) <= 0) {
system.arraycopy(src, low, dest, destlow, length);
return;
}
// merge sorted halves (now in src) into dest
for(int i = destlow, p = low, q = mid; i < desthigh; i++) {
if (q >= high || p < mid && ((comparable)src[p]).compareto(src[q])<=0)
dest[i] = src[p++];
else
dest[i] = src[q++];
}
}