Question

我有'n'行数据集，每个数据集有两个由空格分隔的组件。首先是卡号，第二个是名称。一个人是相同的，如果他有相同的卡号或名称。如何从数据集中找到唯一人的总数？

示例：

1 A

1 B

2 B

3 C

此数据集有2个独特的人。这是因为第一和第二行卡号相同，第二和第三行名称相同。

可以使用哪种算法来解决这类问题？

Answer 1

这是使用图论和连接组件的另一种解决方案：

              int CountUnique(Person[] persons)
                Dictionary<string, int> phones = new Dictionary<string, int>();
                Dictionary<string, int> emails = new Dictionary<string, int>();
                bool[] unique = new bool[n];
                int count = 0;
                int max = 0;
                for (int i = 0; i < n; i++)
                {
                    Person p = persons[i];
                    int pA = -1, pB = -1;
                    if (phones.ContainsKey(p.Phone))
                    {
                        pA = phones[p.Phone];
                    }
                    if (emails.ContainsKey(p.Email))
                    {
                        pB = emails[p.Email];
                    }
                    if (pA != -1)
                    {
                        persons[pA].Next.Add(p);
                        p.Next.Add(persons[pA]);
                    }
                    else
                    {
                        phones.Add(p.Phone, p.Index);
                    }

                    if (pB != -1 && pB != pA)
                    {
                        persons[pB].Next.Add(p);
                        p.Next.Add(persons[pB]);
                    }
                    if (pB == -1)
                    {
                        emails.Add(p.Email, p.Index);
                    }
                }


                int current = 0;
                Person pCurrent;
                count = 0;
                while ((pCurrent = FindUnvisited(persons, current)) != null)
                {
                    BFS(pCurrent);
                    count++;
                }

                return count;


            }

            private static void DFS(Person pCurrent)
            {
                pCurrent.Visited = true;
                foreach (Person p in pCurrent.Next)
                {
                    if (!p.Visited)
                    {
                        BFS(p);
                    }
                }
            }

            private static Person FindUnvisited(Person[] persons, int current)
            {
                for (int i = current; i < persons.Length; i++)
                {
                    if (persons[i].Visited == false) return persons[i];
                }
                return null;
            }


        }
    }

Answer 2

我提出的解决方案是使用某种分区：大多数操作都是在O（1）或O（logn）中完成的，并且由用户完成一次，因此时间复杂度约为O（n），O （logn）取决于Dictionary的实现方式。

int CountUnique(Person persons)
{
        Dictionary<string, int> phones = new Dictionary<string, int>(); //Keep a dictionary where each phone number is mapped to a partition
        Dictionary<string, int> email = new Dictionary<string, int>();  //Keep a dictionary where each email is mapped to a partition
        bool[,] linked = new bool[n, n]; //Lookup table used to tell if 2 partitions are linked (represents the same person)
        int count = 0;
        int max = 0;
        for (int i = 0; i < n; i++)
        {
            Person p = persons[i];
            int pA = -1, pB = -1; // Partition found using the phone number, Partition found using email
            if (phones.ContainsKey(p.Phone))
            {
                pA = phones[p.Phone];
            }
            if (emails.ContainsKey(p.Email))
            {
                pB = emails[p.Email];
            }

            if (pA == -1 && pB == -1) // First case, not found: Add both phones and email and create a new partition. Number of unique persons is also incremented.
            {
                phones.Add(p.Phone.Trim(), max);
                emails.Add(p.Email.Trim().ToLower(), max);
                max++;
                count++;
            }
            else
            {
                if (pA != -1 && pB != -1 && pA != pB) // Found using both parameters on different partitions
                {
                    if (!linked[pA, pB] && !linked[pB, pA]) // If the partition are not linked, link them
                    {
                        count--; // We'lost one partition => one unique person less
                        linked[pA, pB] = linked[pB, pA] = true;
                    }
                }
                if (pA == -1) // We did find an existing email but no phone
                {
                    phones.Add(p.Phone.Trim(), pB); // Add the phone number
                    max++;
                }
                if (pB == -1) // We did find an existing phone but no email
                {
                    emails.Add(p.Email.Trim().ToLower(), pA); // Add the email number
                    max++;
                }
            }

        }



return count;
}

Answer 3

C ++的排序。但把它作为sudo代码。

int uniqueCount = 0;
map<string, bool> column_1;
map<string, bool> column_2;
string left, right
for(int x = 0 ;x < matrix.count; x++) {
   left = matrix[x][0]
   right = matrix[x][1];
   if(column_1.find(left) != column_1.end && column_2.find(right) != column_2.end){
  ++uniqueCount 
   column_1[left] = true;
   column_2[right] = true;
   }
  else --uniqueCount;
}

如果以上内容无法编译，我很抱歉。把它作为sudo代码，我已经有一段时间不在c ++中了，并且不认为rails代码会有所帮助。

如何从具有链接信息的配对数据列表中查找唯一名称的数量？

3 个答案: