How to perform arithmetic operation in pdist?

Hi, i need some advice for solving the error of using squareform,

Assignment has more non-singleton rhs dimensions than non-singleton subscripts
Error in mapmulti (line 234)
            distAllTier1(i,j)   = squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] ));

And the code as follow:

 for i = 1:1:numNodes  
      for j = indPCHTier1 %indPCHTier1 =  1     3     5     7    13    18    21 = j(7)

      if node(i).tier == 1  
          if i == j 
              distAllTier1(i,j)  = NaN; 
          else
              distAllTier1(i,j)   = squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] ));     
          end
      end
      end
  end

Thanks :D

[ANKUR KUMAR]

Please provide us the complete program.

[Asyran Abdullah]

The complete code is too long, here i simplify the value of code.

numNodes  = 100
indPCHTier1 =  1     3     5     7    13    18    21
squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] )) = have Columns 1 through 25 with random number. 

eg:

Then, how can i put the value of "squareform (pdist)" into distAllTier1(i,j) ?

Thanks.

[Asyran Abdullah]

A1  =  pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] );
distAllTier1(i,j)  = squareform(A1);

I try to make something different. Still error.

Here is previous code:

distAllTier1(i,j) = sqrt((node(i).x - node(j).x).^2 + (node(i).y - node(j).y).^2); 

Answer 1

[Walter Roberson]

            distAllTier1{i,j}   = squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] ));

[Asyran Abdullah]

Sir, it came out the cell error.

Cell contents assignment to a non-cell array object.

Error in mapmulti (line 236)
            distAllTier1{i,j}   = squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] ));

[Walter Roberson]

You have a left over distAllTier1 in your workspace that you need to clear. Or else you have an initialization of distAllTier1 that you did not show.

[Asyran Abdullah]

Actually, the distAllTier1 came from this line:

distAllTier1 = zeros(numNodes,numPCHTier1);

which numNodes=100 and numPCHTier1=7. Then distAllTier1 result will be 0 in 100 lines as follow:

>> length(distAllTier1)

ans =

     100

I just want to make the,

squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] ))

result, save in distAllTier1(i,j). How can i do it? Any advice?

[Walter Roberson]

distAllTier1 = cell(numNodes,numPCHTier1);

Answer 2

[Image Analyst]

Why not just use pdist2() and not worry about squareform()?

I always use pdist2() and never use pdist().

[Asyran Abdullah]

like this?

pdist2([node(j).x - node(i).x] + [node(i).y - node(j).y] );

[Walter Roberson]

pdist2( [node(i).x(:), node(i).y(:)], [node(j).x(:), node(j).y(:)] )

perhaps.

This assumes that node(i).x(K) and node(i).y(K) together name X, Y coordinates, so node(i).x and node(i).y are a list of corresponding x and y coordinates for a list of distinct points associated with node(i), and that node(j) likewise has a list of distinct points associated with it, and that you want to find the distance from each point in node(i) to each point in node(j)

The result would be a symmetric 2D matrix of distances.

If your situation is instead that node(i).x and node(i).y both form part of multidimensional coordinates for a single multidimensional point node(i) and likewise there is a single multidimensional point designated for node(j) and you want to find the multidimensional distance, then you would need different code (and you would need to explain why you bothered to split multidimensional coordinates into two components.)

[Asyran Abdullah]

It has an error sir,

Subscript indices must either be real positive integers or logicals.

Error in xcuba1 (line 140)
  node(i).dist =  pdist2( [node(i).x(:), node(i).y(:)], [node(j).x(:), node(j).y(:)] )

Then i try this,

    node(i).dist =  squareform( pdist([x, y]) );

[Walter Roberson]

I suspect that you have not defined i or j at that point. Difficult to say without the code to examine.

[Asyran Abdullah]

This is a original code that use single input:

 distAllTier1(i,j)   = sqrt((node(i).x - baseStation.x).^2 + (node(i).y - baseStation.y).^2);

I and J use to find the distance between them. I modify the code to make it able to operate in multiple vector. Then is use this,

 distAllTier1(i,j)   = squareform (pdist([node(j).x - node(i).x] + [node(i).y - node(j).y] ));

I dont know either in multiple vector it need to find distance between i & j or not.

[Image Analyst]

Make it easy to help you. Attach your node variable in a .mat file.