A question regarding matrices

I have a data file that looks like this:

SECTION Graph
Nodes 10
Edges 5
E 1 2 10
E 2 4 2
E 4 3 1
E 9 10 9
E 4 10 6

E 1 2 10 means there is an edge between 1 and 2 and have a cost 10. I would like to read this data from my data file and create an adjacency matrix G using this data where the G(i,j) is the cost of edge(i,j) and it is -1 if there is no edge. Please Help, I am fairly new to matlab.

[Sean de Wolski]

Could you provide the expected results for the file above?

[jana]

G = [0 10 0 0 0 0 0 0 0 0;
     10 0 0 2 0 0 0 0 0 0;
     0 0 0 1 0 0 0 0 0  0;
     0 0 1 0 0 0 0 0 0 6 ;
     0 0 0 0 0 0 0 0 0 0;
     0 0 0 0 0 0 0 0 0 0;
     0 0 0 0 0 0 0 0 0 0;
     0 0 0 0 0 0 0 0 0 0;
     0 0 0 0 0 0 0 0 0 10;
     0 0 0 0 6 0 0 0 10 0;]

[Jos (10584)]

There are some inconsistencies in your question and comment:

• Does "E 1 2 10" mean that G(1,2) will be 10 but also that G(2,1) should be 10?
• And do you want it to be 0 or -1 when there is no edge?

[jana]

yes, "E 1 2 10" mean that G(1,2) will be 10 but also that G(2,1) should be 10

I am sorry, it is -1 when there is no edge.

Answer 1

[Yatin]

Hi,

Below is a working code for reading your specific file.

function adjMatrix =  adjMatrix()
fileID = fopen('adjMat.txt');

tline = fgets(fileID);
count = 1;
nodes = 0;
edges  = 0;
adjMatrix = [];

while(ischar(tline))
    count = count +1;
    tline = fgets(fileID);
    if(tline == -1)
     break;
    end
    tline =  strtrim(tline);

      if(count == 2)
        nodes = str2num(tline(end-1:end)); 
        adjMatrix = zeros(nodes);
      elseif(count == 3)
          edges = str2num(tline(end));
      else
          splitStr = strsplit(tline);
          node1 = str2num(splitStr{2});
          node2 = str2num(splitStr{3});
          cost  = str2num(splitStr{4});

          adjMatrix(node1,node2) = cost;
          adjMatrix(node2,node1) = cost;

      end    
  end
  fclose(fileID);
  adjMatrix(adjMatrix == 0) = -1;
  end

Answer 2

[Vivek Selvam]

Using dlmread():

function G = adjacencyMatrix()

      filename    = 'myFile.txt';
      delimiterIn = ' ';
      startRow    = 1;
      startCol    = 1;
      M           = dlmread(filename,delimiterIn,startRow,startCol);
      numNodes    = M(1,1); 
      G           = -1*ones(numNodes);
      node1       = M(3:end,1);
      node2       = M(3:end,2);
      cost        = M(3:end,3);
      for k = 1:numel(node1)
          G(node1(k),node2(k)) = cost(k);
          G(node2(k),node1(k)) = cost(k);
      end

end

Answer 3

[Yannick]

Here's another option using TEXTSCAN. Note also the illustration of SPARSE and FULL at the end of the code. The main difficulty here may be to understand how to use cell arrays in MATLAB. You can find more information here: Cell arrays doc. At the very end, I also use logical indexing: Logical indexing doc page

    % Open file
    fileID = fopen('adjMat.txt');

    % Read number of nodes and edges
    H = textscan(fileID, '%s %f', 2, 'HeaderLines', 1);
    disp(['Number of ' H{1}{1} ':'])
    NumberOfNodes = H{2}(1)
    disp(['Number of ' H{1}{2} ':'])
    NumberOfEdges = H{2}(2)

    % Read all remaining lines ("E x x x" lines)
    C = textscan(fileID, '%c %f %f %f');
    ArrayOfInitialNodes = C{2}'
    ArrayOfFinalNodes   = C{3}'
    ArrayOfEdgeWeights  = C{4}'

    % Close file
    fclose(fileID);

    % Create adjacency matrix in sparse format with zeros instead of -1
    % (memory efficient if there are lots of zeros)
    AdjMatrixSparseWithZeros = ...
        sparse(ArrayOfInitialNodes, ArrayOfFinalNodes, ArrayOfEdgeWeights, ...
        NumberOfNodes, NumberOfNodes);
    AdjMatrixSparseWithZeros = ...
        AdjMatrixSparseWithZeros + AdjMatrixSparseWithZeros'

    % Transform it to full format (no compact storage in memory)
    % and replace zeros with -1
    AdjMatrixFullWithNegOnes = full(AdjMatrixSparseWithZeros);
    AdjMatrixFullWithNegOnes(AdjMatrixFullWithNegOnes==0) = -1

Hope this helps provide yet another perspective!