Find column where one value is NaN

Hi,

I have a long matrix containing 2 rows and many thousands of columns. The values are either 0.5 or NaN. I want to be able to locate any columns where only the second row is equal to NaN and the first row is equal to 0.5. For example the matrix:

a = [0.5, NaN, 0.5, 0.5, 0.5;
       0.5, NaN, 0.5, NaN, NaN]

should return the indices [4, 5].

Is this possible using some variation of the find function? I can use find to locate the columns where both rows are equal to either 0.5 or NaN but I can't figure out how to do it where one row is 0.5 and one row isn't. If it isn't possible using find what is the most succint way of achieving this?

Thanks,

Martin

Answer 1

[Stephan]

res = find(~isnan(a(1,:)) & isnan(a(2,:))==1)

Answer 2

[KALYAN ACHARJYA]

"Find column where one value is NaN"

a=[0.5, NaN, 0.5, 0.5, 0.5;
   0.5, NaN, 0.5, NaN, NaN];
[r,c]=find(isnan(a));
col_data=setdiff(c,r)