Thread Subject: If there is a NaN anywhere in a 4d matrix, I want to remove the column. How?

I have a 4d matrix that has many columns that are NaNs. I want to remove all of these columns.

So simplified, it's something like this...

y(:,:,1,1) =
     1 2 NaN 4
     1 2 NaN 4
     1 2 NaN 4
     1 2 NaN 4

y(:,:,2,1) =
     1 2 3 4
     1 2 3 4
     1 2 3 4
     1 2 3 4

y(:,:,1,2) =
     1 NaN 3 4
     1 NaN 3 4
     1 NaN 3 4
     1 NaN 3 4

y(:,:,2,2) =
     1 2 3 4
     1 2 3 4
     1 2 3 4
     1 2 3 4


I tried...
>> for a=1:4; for b=1:2; for c=1:2; if isnan(y(1,a, b, c)); y(:,a,:,:)=[]; end; end; end; end

but I wind up with...

??? Index exceeds matrix dimensions.
 >> y

y(:,:,1,1) =
     1 NaN 4
     1 NaN 4
     1 NaN 4
     1 NaN 4

y(:,:,2,1) =
     1 3 4
     1 3 4
     1 3 4
     1 3 4

y(:,:,1,2) =
     1 3 4
     1 3 4
     1 3 4
     1 3 4

y(:,:,2,2) =
     1 3 4
     1 3 4
     1 3 4
     1 3 4


What I really want is...

y(:,:,1,1) =
     1 4
     1 4
     1 4
     1 4

y(:,:,2,1) =
     1 4
     1 4
     1 4
     1 4

y(:,:,1,2) =
     1 4
     1 4
     1 4
     1 4

y(:,:,2,2) =
     1 4
     1 4
     1 4
     1 4


Please help me with this issue?

Thanks so much!
John

 y(:,any(any(any(isnan(y),1),3),4),:,:) = []

% Bruno

Wonderful.

Thanks Bruno!

Best,
John

I spoke a little too soon.

That certainly does the trick with what I gave you, but I need it to work for a variable number of columns and rows. In my data, I have anywhere from 10-24 columns and 60-150 rows.

I'm picking apart the command you gave me now to see if I can get it, but no success yet.

Thanks for the help,
John



"Bruno Luong" <b.luong@fogale.findmycountry> wrote in message <jglirn$18l$1@newscl01ah.mathworks.com>...
> y(:,any(any(any(isnan(y),1),3),4),:,:) = []
>
> % Bruno