Multiply specific parts of cell with cell column

How can I multiply specific parts of a cell array with a cell column?

e.g. y is a 3x5 cell

10 11 12 13 14 
15 16 17 18 19 
20 21 22 23 24

I want to multiply the y(2:3,2:3) party with y(2:3,5)

16 17 * 19 
21 22   24

result should a 2x2 array:

16*19 17*19
21*14 22*14

My attempt did not work:

x = zeros(3,2); %create new empty array (first line should be empty)
 for i=2:3
     x(i,:)=cell2mat(y(i,2:5)).*cell2mat(y(i,5))); 
 end

Answer 1

[Adam]

Why do you have a cell array in the first place instead of a numeric array? It is far slower and requires you to use much uglier constructs to do simple maths operations.

y(2:3,2:3) = y(2:3,2:3) .* y(2:3,5);

would work on a numeric array (in R2016b - Andrei Bobrov's alternative is what I always used pre-R2016b)

y = cell2mat( y );

before that will allow you to get away from your cell array.

Answer 2

[Andrei Bobrov]

y1 = {10 11 12 13 14 
15 16 17 18 19 
20 21 22 23 24};

y = cell2mat(y1);
out = y(2:3,2:3).*y(2:3,5);               % R2016b and Octave
out = bsxfun(@times,y(2:3,2:3),y(2:3,5)); % R2016a and before

Answer 3

[Alexandra Harkai]

There could be a typo here:

cell2mat(y(i,2:5))

Probably should be:

cell2mat(y(i,2:3))

So result should be:

16*19 17*19
21*24 22*24

This can be achieved like this, without a loop (see https://uk.mathworks.com/help/matlab/ref/bsxfun.html):

x = zeros(3,2); %create new empty array (first line should be empty)
x(2:3,:) = bsxfun(@times, cell2mat(y(2:3,2:3)), cell2mat(y(2:3,5)));

Answer 4

[Ki Loius]

Thank you for your answers. I just recognized, that I did not provide enough information for my case. The first row and line in my cell are strings. (first two answers do not work)

Answer 3 is not working, since Subscripted assignment dimension mismatch. I have to multiply a NxM part of the matrix with one column

a more precise example would be cell y:

i a  b  c  d  e
f 11 12 13 14 15
g 16 17 18 19 20
h 21 22 23 24 25
i 26 27 28 29 30

I want to multiply y(2:4,2:3) * y(2:4,6)

[Alexandra Harkai]

Since the example x array was initialised as an all-0 array, it looked like you wanted a numerical array. If that's not the case, you can always do num2cell(...) before the assignment.

[Adam]

It's generally a very bad idea to have data structured in that way with a row and column like that (a table structure sounds more obvious if that is what you want), but you can still apply any of the solutions given with a bit of alteration.

Just pull out the part of the array you want and apply to that, then re-insert it, using num2cell and cell2mat either side of your operation and depending whether you have R2016b or not use bsxfun to do the multiplication.

Answer 5

[Ki Loius]

I think I have another problem:

y is a 17x9 cell

x=cell2mat(y(2:3,2:5)) gives me the content of the cell in that range

but

x=cell2mat(y(2:4,2:5)) or x=cell2mat(y(2:3,2:6)) does not work. It says

Dimensions of matrices being concatenated are not consistent.

[Alexandra Harkai]

This could be a separate Question.

Answer 6

[Ki Loius]

I could solve my problem.

I had empty cells in my cell array. I filled them with 0 dann answer 3 worked.

Thank you!