How to take an average of row elements in a two-column matrix when both elements do not contain a NaN, but use only the non-NaN value when there is a NaN?

I have a matrix with two columns of data. The data are measurements collected at the same times from different sensors, TT1 and TT2. I want to take an average of each row. My problem is dealing with NaN values. The column of data from sensor TDT2 contains many NaNs.

I know that I can use "nanmean" to ignore the rows where Column TDT2 contains a NaN. However, what I want to do is use only data from sensor TT1 (column 1) when there is a NaN from sensor 2 (column 2). In other words, I want to ignore the existence of column 2 when it contains a NaN.

Unfortunately, I can't figure out how to do this. If anyone has any suggestions, I would be very grateful. Thank you.

Here is an example (below) of a starting matrix and what I would want to end up with as a result:

2   2           --> 2 
4   NaN             4
9   1               5

Answer 1

[Bruno Luong]

>> A=[2 2; 4 NaN; 9 1]

A =

     2     2
     4   NaN
     9     1

>> mean(A,2,'omitnan')

ans =

     2
     4
     5