find a Nan in column of cell array and delete the row

Hi everyone, I have a cell array (1152,4) In this cell array I would like see if in each column 1,2,3,4 if there is a Nan, if there is a Nan I would like delete the row.

For that I try to use : cellfun(@isnan,newTab,'UniformOutput',false) and it sends a matrix with logical value (0/1), how can I delete row if there is a 1 in my matrix?

Thank you

Answer 1

[sixwwwwww]

here is an example which can give you idea how you can do it:

a = rand(1152, 4);
a(randi(1152, 1, 20), :) = NaN;
a = num2cell(a);
b = cellfun(@isnan, a);
idx = find(b(:,1));
for i = 2:size(a, 2)
    idx = union(idx, find(b(:,i)));
end
a(idx, :) = [];

Answer 2

[Alex]

thank you, just why you start the loop to i=2 and not i=1?

[sixwwwwww]

because of this:

idx = find(b(:,1)); and also we need to use intersect afterwards which need to arrays so i calculated the first and then started the loop from 2

Answer 3

[Alex]

and why the loop is used ? If I try only:

b = cellfun(@isnan, a);
 idx = find(b(:,1));
 a(idx, :) = [];

It works?!?

[sixwwwwww]

no it will not work because if NaN will appear in 2nd or 3rd or 4th column then that row will not be deleted because we are not getting indices for those rows. I edited my answer which has small mistake. So you can accept my edited question. I correct it

Answer 4

[Jos (10584)]

If your cell array C has only numbers or NaNs, this will do the job

C = {1 2 3 4 ; 11 12 NaN 14 ; 21 22 23 24}
C(any(isnan(cell2mat(C)),2),:) = []

If your cell array C has a mixture of numbers, char arrays or NaNs, the one-liner above will not work, but this will:

C = {1 2 3 4 ; 'DeleteThisRow' 12 NaN ones(3) ; 21 rand(2) 23 24:29}
C(any(cellfun(@(x) numel(x)==1 & isnumeric(x) && isnan(x),C),2),:) = []

Answer 5

[Alex]

thank you but my real probleme is that I got several table:

infrastructure.capteur(1,1).tableau %1st tab
infrastructure.capteur(2,1).tableau %2nd tab
... there a 8 tab

When there is a Nan is a row, i need to delete each rows in each tabs

[sixwwwwww]

then you can use for loop to do it. Access data from each tab in a big cell array and then check each cell array and delete the rows which contain NaN