Hashmap representation of graph

Question

I have a text file with graph edges, for example

1 2

1 3

2 5

etc. , and want to represent my graph in some way. I tried to use a hashmap, is it the best way to represent edges? And second question, how can I access first and second entries in my hashmap? My code here

    DataInputStream dStream = new DataInputStream(new FileInputStream("C:/Programming/Java/test.txt"));
    BufferedReader bReader = new BufferedReader(new InputStreamReader(dStream));

    HashMap<Integer, Integer> graphEdges = new HashMap<Integer, Integer>();
    String line;
    while( (line = bReader.readLine()) != null) {
        String[] firstSecond = line.split(" ");
        int firstDigit = Integer.parseInt(firstSecond[0]);
        int secondDigit = Integer.parseInt(firstSecond[1]);

        graphEdges.put(firstDigit, secondDigit);
    }

    System.out.println(graphEdges);

    bReader.close();
}

Answer 1

Among many possible representations of a graph, the 2 basic are the following:

• Adjacency lists

In Java:

Map<Integer, List<Integer>> graph = new HashMap<>();
...
while( (line = bReader.readLine()) != null) {
    String[] tokens = line.split(" ");
    int firstNode = Integer.parseInt(tokens[0]);
    int secondNode = Integer.parseInt(tokens[1]);

    if(!graph.containsKey(firstNode)) graph.put(firstNode, new LinkedList());
    graphEdges.get(firstNode).add(secondNode);
}

• Adjacency Matrix

In Java:

int[][] graph = new int[numberOfNodes][numberOfNodes];
while( (line = bReader.readLine()) != null) {
    String[] tokens = line.split(" ");
    int firstNode = Integer.parseInt(tokens[0]);
    int secondNode = Integer.parseInt(tokens[1]);

    graph[firstNode-1][secondNode-1] = 1;
    graph[secondNode-1][firstNode-1] = 1;
}

Below is a comparison of operations and storage efficiency of these 2 representations:

                   |     Adjacency List    |   Adjacency Matrix   |
Storage            |      O(nodes+edges)   |     O(nodes^2)       |
Add node           |          O(1)         |     O(nodes^2)*      |
Add edge           |          O(1)         |        O(1)          |
Remove node        |        O(edges)       |     O(nodes^2)*      |
Remove edge        |        O(edges)       |        O(1)          |
isAdjacent(x1,x2)  |        O(nodes)       |        O(1)          |
*Requires copying of the whole array

You can also make a slight modification in the adjacency lists and use HashSets, instead of LinkedList for storing the adjacent nodes. In that case, everything is the same, except the isAdjacent(x1,x2) operation that now has O(1) complexity (amortized).