How to remove NaNs from rows in matrices within a cell array?

Hi all,

I have this cell array participant _H where some of the matrices inside the cells include NaNs. When I try and remove all of the NaNs using cellfun, like so

participant_H(cellfun(@(participant_H) any(isnan(participant_H)),participant_H)) = []

I just get an empty cell array (see attachment).

Could someone kindly help?

Answer 1

[the cyclist]

Instead of trying to set the NaN element to [], you can just select the non-NaN elements.

participant_H_no_nan = cellfun(@(x)x(not(isnan(x))),participant_H,'UniformOutput',false);

I typically find it clearer to use a different variable name (in this case x) in the anonymous function.

[lil brain]

Superb! Thanks for the answer!

[lil brain]

@the cyclist could you be so kind to also tell me how I can remove the rows with NaNs in the matrices within the cells of my array called participant _columns (see attachment)? Note that in reality my array contains much more cells that just this one.

I have tried using the code you supplied above but it only leaves the first column and removes the other 20 (see attachment participant_columns_no_nans).

That would be greatly appreciated. Thanks!

Note: I

[the cyclist]

The prior code looked at each vector (inside each cell), and kept only the the non-NaNs.

That doesn't do what you intended when acting on matrices. Unlike what you stated, it is not returning the first column. It is returning all the non-NaNs of the entire matrix, in one long vector.

The following code does what you intend, I believe. Using the all function, it finds the rows where all elements are non-NaN, and returns only those rows.

participant_columns2_no_nan = cellfun(@(x)x(all(not(isnan(x)),2),:),participant_columns2,'UniformOutput',false);

I feel that I am "fishing" for you here, and not "teaching you to fish". I strongly suggest that you really try to understand what both of these pieces of code are doing, and not just use it blindly.

I would suggest you just try out on a single matrix, and not worry about the cellfun complication at first. See what each piece does, and understand it.

% Define an input
A = [ 1  2  3  NaN;
      4  5  6   7;
      8  9 10  11;
     12 13 14  NaN];

% Which elements are NaN?
isnan(A)

ans = 4×4 logical array
   0   0   0   1
   0   0   0   0
   0   0   0   0
   0   0   0   1

% Which are not NaN?
not(isnan(A))

ans = 4×4 logical array
   1   1   1   0
   1   1   1   1
   1   1   1   1
   1   1   1   0

% Rows in which all elements are not NaN
all(not(isnan(A)),2)  % Read the documentation of all(), to understand the syntax

ans = 4×1 logical array
   0
   1
   1
   0

% Use the above logical index to index into A, and pull only the desired
% rows
A(all(not(isnan(A)),2),:)

ans = 2×4

4 5 6 7 8 9 10 11