• [LintCode] Permutations II


    Given a list of numbers with duplicate number in it. Find all unique permutations.

    Example

    For numbers [1,2,2] the unique permutations are:

    [
      [1,2,2],
      [2,1,2],
      [2,2,1]
    ]
    
    Challenge 

    Using recursion to do it is acceptable. If you can do it without recursion, that would be great!

    This problem is a follow up of Permutation with an extra condition: there are duplicated numbers in the given list. 

    In order to avoid duplicated permuations, we use a "representative selection" principle.

    Take [1, 2, 2, 2, 3] as an example,  when the current permutation needs a 2 and there are more than one 2 left in the 

    unused numbers pool, if we use the same algorithm with Permuation(no duplicated numbers), we would have duplicated answers.

    This is shown as follows.

    Say, we've already picked 1 and 2 for the current permutation. {2, 2, 3} are left for further picks. If we need to pick another 2, we can either 

    pick the first or second 2 in {2, 2, 3}, both leaving {2, 3} for further picks. These two picks generate duplicated permutations. To avoid this,

    we rule that we only pick the first unpicked 2 and skip the case where we pick the second 2 out of {2, 2, 3}. 

    To achieve this "representative selection" principle, we need to sort the input array first so that all duplicated numbers are adjacent. This way,

    when picking duplicated numbers, we can check if its previous duplicated neighbor has been picked or not. If it has, we know this is not a duplicate 

    case; If it hasn't, then we know for the same permutations, we could've picked its previous unpicked neighbor as their representative, skip this case.

     1 public class Solution {
     2     public List<List<Integer>> permuteUnique(int[] nums) {
     3         List<List<Integer>> results = new ArrayList<>();
     4         if(nums == null || nums.length == 0) {
     5             results.add(new ArrayList<Integer>());
     6             return results;
     7         }
     8         boolean[] used = new boolean[nums.length];
     9         for(int i = 0; i < used.length; i++) {
    10             used[i] = false;
    11         }
    12         Arrays.sort(nums);
    13         permuteUniqueDfs(results, new ArrayList<Integer>(), nums, used);
    14         return results;
    15     }
    16     private void permuteUniqueDfs(List<List<Integer>> results, List<Integer> list,
    17                                     int[] nums, boolean[] used) {
    18         if(list.size() == nums.length) {
    19             results.add(new ArrayList<Integer>(list));
    20             return;
    21         }
    22         for(int i = 0; i < nums.length; i++) {
    23             if(used[i] || (i > 0 && nums[i] == nums[i - 1] && used[i - 1] == false)) {
    24                 continue;
    25             }
    26             list.add(nums[i]);
    27             used[i] = true;
    28             permuteUniqueDfs(results, list, nums, used);
    29             list.remove(list.size() - 1);
    30             used[i] = false;
    31         }
    32     }
    33 }

    Related Problems

    Next Permutation II

    Permutation Sequence

    Next Permutation

    Permutations

  • 相关阅读:
    Hadoop 2.6.0-cdh5.4.0集群环境搭建和Apache-Hive、Sqoop的安装
    Hive数据仓库笔记(三)
    Hive数据仓库笔记(二)
    Hive数据仓库笔记(一)
    HBase新的客户端接口
    Hadoop之Secondary NameNode
    hadoop2.x源码编译
    基于新浪SAE平台的微信开发
    分布式存储系统-HBASE
    编程珠玑--心得
  • 原文地址:https://www.cnblogs.com/lz87/p/7494138.html
Copyright © 2020-2023  润新知