From: Walter Roberson on
joseph Frank wrote:
> SampleAll 478x221 array. it si a mix of nan,numbers and strings but
> the third column is only numbers and nans

Okay, that eliminates issues with nesting, but you did not clarify what output
you are looking for. If you want a 478x221 array of logical, with (J,K) being
true of SampleAll{J,K} contains a nan, then the cellfun with any(isnan(V))
will do. If you want the indices of where the nan are within SampleAll{J,K}
then more work would be required.
From: joseph Frank on
unfortunately I need the indices
From: Jan Simon on
Dear Joseph!

> unfortunately I need the indices

Although I do love CELLFUN, especially for the very fast builtin commands, it might be a good idea to start with creating a simple double loop here:
A = cell(m, n); <your data>
[nm, nn] = size(A);
for im = 1:nm
for in = 1:nn
NaNIndex = find(isnan(A{im, in}));
<perform your operations on the NaN elements here>
end
end
This might be slower or faster than the CELLFUN approach. It is not as nice as CELLFUN, but simple.

Kind regards, Jan