For elements in two arrays/columns of numbers, how can I find the average only if both elements are non-zero?

I have many external files. Each file contains two long columns of numbers, like this:

1        2
11       12
6        48
9        0.0
3        0.5

I can read each file in. Now I want to calculate the average of each row. In the column of results, for example, the elements would be:

(1 + 2 / 2) 
(11 + 12 / 2)
(6 + 48 / 2) ...

The problem is, I don't want to take the average of rows where one element is zero. For example, I want the row that has

9        0.0

to be "9." Does anyone know how I can do, "calculate the average of each row, but leave the row alone if one element is zero?" I'm using 2016a. Thank you for any advice.

Answer 1

[madhan ravi]

matrix(matrix == 0) = nan
nanmean(matrix, 2)

[Srh Fwl]

@madhan ravi, thank you very much. Does this mean, if there's a zero OR a nan element, ignore it? That would be helpful. I do have nan elements too. Thanks.

[madhan ravi]

Or

 cellfun(@(x) mean(x(x ~= 0)), num2cell(matrix, 2))

[madhan ravi]

Yes, your right.

[madhan ravi]

 cellfun(@(x) mean(x((x ~= 0) & (~isnan(x)))), num2cell(matrix, 2))

[Srh Fwl]

Thank you, @madhan ravi. I found that I couldn't use "nanmean" because I don't have the statistics toolbox. However, I found an alternative and it works with the first part of your suggestion. Really appreciate it.

A = [3; 4; 5; 0]
B = [7; 8; 9; 10]
C = [A(:), B(:)]
C(C == 0) = nan
D = mean(C,2,'omitnan')