Find row with NaN in it.

I have variable X.

 X=[2   4   6;
    4   5   NaN
    3   NaN NaN
    5   8   12
    3   8   NaN];

How to find any row with NaN in it. ismember([3 8 NaN],X,'rows') gives me 0.

[Artyom]

I mean how to find specific row in the matrix (for example 'find' [3 8 NaN] must return me [0;0;0;0;1];).

Answer 1

[per isakson]

A start:

    is(:,1) = X(:,1)==3;
    is(:,2) = X(:,2)==8;
    is(:,3) = isnan( X(:,3));
    all( is, 2 )

which returns

    ans =
         0
         0
         0
         0
         1

AFAIK: isnan is the only way to spot NaN.

I don't think there is a magic one-liner. A little function along the example above is one way. Loop over all columns, ....

Answer 2

[Guillaume]

[rows, columns] = find(isnan(X));
unique(row);

will give you the row indices

[sworland]

simple and fast, 1 upvote

Answer 3

[Roger Stafford]

The code

   t = any(isnan(X),2);

will return you a column vector of logicals in which each element is true if the corresponding row of X contains a NaN. I trust that would meet your needs. In your example you would get t = [false;true;true;false;true].

Answer 4

[Andrei Bobrov]

   X=[2   4   6;
      4   5   NaN
      3   NaN NaN
      5   8   12
      3   8   NaN];
  с = [3 8 NaN];

z = [X;c];
z(isnan(z)) = max(z(:)) + 1;
out = ismember(z(1:end-1,:),z(end,:),'rows');

other example

X = randi(100,200,15);
X([5,125],[2,10]) = nan; % your array
c = X(5,:);

z = [X;c];
z(isnan(z)) = max(z(:)) + 1;
out = ismember(z(1:end-1,:),z(end,:),'rows');