关于将项目添加到自定义数据类型列表的问题

Question

I have a simple node class. Now I want to add all the adjacent neighbors to every node in a list I have, but for some reason it just adds the first adjacent node of the neighborhood. In the below code "edges" is a datatree( basically a dictionary-> int,List) where every consists of to integers representing a node. Maybe someone could explain me why It just adds one value to the list of every node, I would be really grateful!

The node class

public class Node
{
  //Properties
  public int name;
  public List<Node> AdjN;
  
  // constructor that takes the current node and initializes the adjList
  public Node(int _node)
  {
    name = _node;
    AdjN = new List<Node>();
  }
}

The main code:

   var nodes = new List<Node>();   
    for (int i = 0; i < edges.BranchCount; i++) // edges is a datatree which is basically a dictionary (int,List<int>
    {
      Node n1 = new Node(edges.Branch(i)[0]);// this int is the first node of the current edge
      Node n2 = new Node(edges.Branch(i)[1]);/ this int is the second node of the current edge
      
      nodes.Add(n1);
      nodes.Add(n2);
      n1.AdjN.Add(n2); 
      n2.AdjN.Add(n1);     
    }
 
    
    var neighbors = new List<int>();
    for (int i = 0; i < nodes[0].AdjN.Count; i++)
      neighbors.Add(nodes[0].AdjN[i].name);

Thanks everybody for looking at this!

Answer 1

Try following:

using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Net;
using System.Net.Sockets;

namespace ConsoleApplication1
{
    class Program
    {
        static void Main(string[] args)
        {
            Dictionary<int, List<int>> edges = new Dictionary<int,List<int>>() {
                { 0, new List<int>(){100,200}},
                { 1, new List<int>(){300,400}},
                { 2, new List<int>(){500,600}},
                { 3, new List<int>(){700,800}}
            };

            var nodes = new List<Node>();   
            for (int i = 0; i < edges.Count; i++) // edges is a datatree which is basically a dictionary (int,List<int>
            {
              Node n1 = new Node(edges[i][0]);// this int is the first node of the current edge
              Node n2 = new Node(edges[i][1]);// this int is the second node of the current edge
      
              nodes.Add(n1);
              nodes.Add(n2);
              n1.AdjN.Add(n2); 
              n2.AdjN.Add(n1);     
            }
 
    
            var neighbors = new List<int>();
            for (int i = 0; i < nodes[0].AdjN.Count; i++)
              neighbors.Add(nodes[0].AdjN[i].name);
  
        }
    }
    public class Node
    {
        //Properties
        public int name;
        public List<Node> AdjN;

        // constructor that takes the current node and initializes the adjList
        public Node(int _node)
        {
            name = _node;
            AdjN = new List<Node>();
        }
    }
}

Answer 2

The problem with your current implementation is that a node will appear in your list multiple times if it is connected to multiple edges, each time with one neighbor. It should instead appear in your list only once, with one neighbor per edge. Here's an example graph:

B -- A -- C

Currently, you process it as follows:

1. Add nodes B & A to the list, each listing the other as a neighbor:
   • B, neighbors: A
   • A, neighbors: B
2. Add nodes A (again) & C to the list, each listing the other as a neighbor.
   • B, neighbors: A
   • A, neighbors: B
   • A, neighbors: C
   • C, neighbors: A

Node A appears twice in your list, once listing B as a neighbor and once listing C. The second loop iteration ought to notice that A is already in the list, and add C as another neighbor.

But the only way to see whether A is already in your List is to traverse the entire thing, which is not efficient when the number of nodes is large.

Try changing your nodes variable from a List<Node> data structure to a Dictionary<int,Node> , where the key is the node identifier and the value is the complete node. On each iteration through the loop, you can check for the existence of n1 and n2 in your dictionary in O(1) time, and add them to the dictionary only if they aren't yet present.

Don't forget that if you find the node identifier in your Dictionary , you need to update the neighbor list for the object in your dictionary, not the object you created and just happens to have the same identifier.

The algorithm should look something like this:

For each edge
    For j = 0 to 1
        id[j] = node[j] attached to the edge
        if id[j] is in the dictionary
            set n[j] = value from dictionary
        else
            set n[j] = new Node(id[j])
            add n[j] to the dictionary

    add n[1] as neighbor to n[0]
    add n[0] as neighbor to n[1]