Deleting the row of a 3D matrix with a condition

I have a simple 3D matrix. Every time I do this operation on the (:,:,1) matirx it gives me an error saying "A null assignment can have only one non-colon index." I dont understand this. But in a simple 2D matrix it works. If I assign "NaN" instead of "[]" then the second row is assigned with NaN NaN values (in 3D). Can I not delete the row in a 3D matrix? Appreciate some insight to this.

A(:,:,1) = [1 2 ; 0 5 ; 7 8 ];
A(:,:,2) = [10 11 ; 1 14 ; 0 17 ];

A((A(:,1,1).*A(:,2,1))==0,:,1)=[]

[KALYAN ACHARJYA]

Look here you will get the exact answer

As per undestanding from your question, you have to delete the complete plane, not a row line.

Answer 1

[Jan]

Of course you cannot delete one row in a 3D array: The resulting array must have the same length for all subvectors. Removing a row from a 3D array is equivalent to removing one element from a matrix: The result would not be rectangular anymore, but matrices must have the same number of elements for all rows, and all columns, respectively.

A(:,:,1) = [1 2 ; 0 5 ; 7 8 ];
A(:,:,2) = [10 11 ; 1 14 ; 0 17 ];
B = A;
B(:, :, 1) = []  % Working
A(1, :, 1) = []  % Error: output cannot be a valid array

Therefore I cannot guess, what you expect as output and cannot suggest a solution. Assigning NaN to a row works, because this does not change the number of elements.

[Ashane Fernando]

Hi Jan,

Thank you very much for your answer. I want A(:,:,1) to be [1 2; 7 8] and A(:,:,2) to be [10 11; 1 14] respectively.

Is this possible?

[Jan]

@Ashane: Now we have a not working code and one example for the wanted operation. It is more reliable if you explain the wanted operation in addition.

What is the wanted output for e.g.:

A(:,:,1) = [1 2 ; 0 5 ; 7 8 ];
A(:,:,2) = [10 11 ; 1 14 ; 2 17 ];

when A contains one 0 only?

[Ashane Fernando]

Jan, then A(:,:,2) is actually,

A(:,:,2) = [10 11; 1 14; 0 17];

onece the columns multiply (e.g. A(:,:,1) the answers are 2, 0 and 56. I do not want the 0s. After I want A(:,:,1) = [1 2; 7 8] to belike that. I guess it is not possible in a 3D array. Thus I got some "NaN"s to cover up for the rest.

[Jan]

Sorry, I cannot follow you. I asked you what happens if you have only 1 zero in the input, such that you cannot remove a single subvector without destroying the rectangular shape of the array. Now your answer is: If the array contains one 0 only, then it actually contains two? What does this mean?

If all submatrices contain the same number of zeros, than you can remove the corresponding rows or columns or what you want. But you did not clarify yet, if this is guaranteed.

[Ashane Fernando]

Hi Jan,

I realised I was looking at a wrong thing. I figured that I cannot delete a row from a 3d matrix.

I converetd the necessary values to a 2D matrix and did the manipulation as I required.

Thanks for showing me some light.