Azzera filtri
Azzera filtri

How to remove Nans from matrix

2 visualizzazioni (ultimi 30 giorni)
Ian
Ian il 10 Set 2013
I have a matrix (size 123317x6), and every 128 rows I have 3 nans on the last three columns of that thoses rows. Now what I want to do is remove all the NaNs completely without removing that row or column (just the NanNs) The data on the next row then needs to be moved up to where the Nans were. I will give you a quick example.
so I want this matrix:-
M=
1 3 6 nan nan nan 4 5 4 1 3 4 1 2 2 6 3 1 1 3 4 nan nan nan 3 4 5 6 7 8
M=
1 3 6 4 5 4 1 3 4 1 2 2 6 3 1 1 3 4 3 4 5 6 7 8
Hope this example explains what I want better!
Any help would be greatly appreicated.

Risposta accettata

Azzi Abdelmalek
Azzi Abdelmalek il 10 Set 2013
Modificato: Azzi Abdelmalek il 10 Set 2013
out=reshape(M(~isnan(M)),[],size(M,2))
Edit
M can not be reshaped unless we add some nan at the end of M
%---------Your array-------------------
M=rand(123317,6);
M(128:128:n,end-2:end)=nan;
%-----------------------------------------
n=size(M,1);
M=M(~isnan(M));
M(end:end+mod(6-mod(numel(M),6),6))=nan;
M=reshape(M,[],6);
  1 Commento
Geert
Geert il 10 Set 2013
I like your answer more than mine. There is however a small mistake, it should be:
out=reshape(M(~isnan(M)),[],size(M,2)-3)

Accedi per commentare.

Più risposte (3)

Andrei Bobrov
Andrei Bobrov il 10 Set 2013
Modificato: Andrei Bobrov il 10 Set 2013
out = resM(M==M)
ADD
M1 = M';
out = reshape(M1(M1==M1),size(M,2),[])';

Geert
Geert il 10 Set 2013
If it are exactly 3 NaN's per row, you could do something like this:
M = [1 3 6 nan nan nan 4 5 4 1 3 4 1 2 2;
6 3 1 1 3 4 nan nan nan 3 4 5 6 7 8];
M_new = zeros(size(M,1), size(M,2)-3);
for ii = 1:size(M,1)
M_new(ii,:) = M(ii,~isnan( M(ii,:)));
end

Ian
Ian il 15 Set 2013
This for you help guys much appreciated! :)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by