For command in function

I have a function attached and would like it to automatically run it with x = 2:4 & y = 2:4, that is, the function to run 8 times & have 8 outputs, with inputs

x =2, y = 2

x = 2, y = 3

x = 2, y = 4

x = 3, y = 2

...

Should I write a 'for' command with the function? It does not seem to work the best this way.

[Stephen23]

"It does not seem to work the best this way. "

You could use a loop, but it is often simpler and more efficient to write vectorized code:

https://www.mathworks.com/help/matlab/matlab_prog/vectorization.html

[Rik]

And if you vectorize it, you can use implicit expansion, or use ndgrid to generate all combinations.

[Cside]

this is the current vectorized code I tried to create:

[x,y] = ndgrid (2:4,2:4)

%%vectorized example

xy = rowsncolumns (:,1) ==x & rowsncolumns(:,2) == y;

result = mean (dataset_pfc_tar_57_n1(xy))

But the error was that matrix dimensions must agree, and an error in line 3. How should i change the ndgrid to fit the vectorized code?

FYI

rowsncolumns = 621 x 2 double

[Walter Roberson]

You are attempting to compare a 621 x 1 vector to a 3 x 3 array output by ndgrid

[Cside]

my inputs of x & y are meant to be scalar though, i.e. x = 2, y = 2. What could i write to allow the vectorized code to run x = 2:4 & y = 2:4?

[darova]

What is length of dataset_pfc_tar_57_n1?

[Cside]

1 x 621 double

[darova]

Each value of (x,y) you want to compare to rowsncolumn?

As KALYAN ACHARJYA said

>> combvec([2:4],[2:4])
ans =
     2     3     4     2     3     4     2     3     4
     2     2     2     3     3     3     4     4     4     

If i understood you correctly output matrix should be of size 621 x 9 ?

[Cside]

yes, for each value, i would compare it to rowsncolumn. the output at the end should only give me 9 x 1, or 1x 9, basically 9 values as the last line of code is the mean.

[Walter Roberson]

After you use ndgrid to build x y you can

x = permute(x, [3 1 2])

and the same for y. Then assuming R2016b or later, the & would give a 3d result, 621 by 3 by 3

Now take that and .* by dataset and sum along the first dimension. Then sum the result of the & itself along the first dimension to get counts. ./ the two qualities. You should now have a 1 x 3 x 3 array of means.

Answer 1

[darova]

You can use pdist2()

[X,Y] = meshgrid(2:4);
x = X(:);
y = Y(:);
D = pdist2(rowsncolumns,[x y]);
% D(:,1) - first column of D - distance from (rowscolumns) to (x1,y1) point 

Indices you want: D == 0

Size of D: 621 x 9

[Walter Roberson]

Ah, but how do you proceed from here to vectorize the mean() that has to be done for each x y pair?

[darova]

Maybe not to use mean()

ix = D == 0;
data = repmat(dataset_pfc_tar_57_n1,[1 9]);
result = sum(data.*ix) ./ sum(ix);

size of result is 1 x 9

[Cside]

thank you so much, I really appreaciate everyone's help :D