算法总结之 在两个排序数组中找到第K小的数
给定两个有序数组arr1 和 arr2 ,再给定一个int K,返回所有的数中第K小的数
要求长度如果分别为 N M,时间复杂度O(log(min{M,N}),额外空间复杂度O(1)
解决此题的方法跟之前的求两个数组求中位数的情况,如出一辙~ 非常给力!
此题目需要分情况讨论:
假设长度较短的数组长度 lenS 较长的lenL
情况1、 K<1 或者 K>lenS+lenL k值无效
情况2、 k<=lenS 分别在两数组选择第前 k个数, 然后取其中位数
情况3、 k>lenL
package TT; public class Test13 { public static int getUpMedian(int[] a1, int s1, int e1,int[] a2, int s2, int e2){ int mid1 = 0; int mid2 =0; int offset = 0; while(s1<e1){ mid1 = (s1+e1) /2; mid2 = (s2+e2) /2; offset = ((e1-s1+1)&1)^1; if (a1[mid1]>a2[mid2]) { e1 = mid1; s2 = mid2+offset; }else if(a1[mid1]<a2[mid2]){ s1 = mid1 + offset; e2= mid2; }else { return a1[mid1]; } } return Math.min(a1[s1], a2[s2]); } public static int findKthNum(int[] arr1, int[] arr2, int kth){ if(arr1==null || arr2==null){ throw new RuntimeException("are you ok?"); } if(kth<1 || kth>arr1.length+arr2.length){ throw new RuntimeException("too long"); } int[] longs = arr1.length >=arr2.length ? arr1 :arr2; int[] shorts = arr1.lengtharr1 :arr2; int l = longs.length; int s = shorts.length; if(kth <= s){ return getUpMedian(shorts, 0, kth-1, longs, 0, kth-1); } if(kth>l){ if(shorts[kth-l-1]>=longs[l-1]){ return shorts[kth-l-1]; } if(longs[kth-s-1]>=shorts[s-1]){ return longs[kth-s-1]; } return getUpMedian(shorts, kth-l, s-1, longs, kth-s, l-1); } if (longs[kth-s-1]>=shorts[s-1]) { return longs[kth -s -1]; } return getUpMedian(shorts, 0, s-1, longs, kth-s, kth-1); } public static void main(String[] args ){ int[] a1 = new int[4]; int[] a2 = new int[4]; a1[0]=0; a1[1]=1; a1[2]=2; a1[3]=3; a2[0]=4; a2[1]=5; a2[2]=5; a2[3]=6; int kth=4; int c = findKthNum( a1, a2, kth); System.out.println(c); } }
结果:
posted @ 2017-08-23 16:13 toov5 阅读( ...) 评论( ...) 编辑 收藏