How can I split two side-by-side vectors into many individual vector pairs?

I have several 10368x1 vectors, of which each of the 10368 pieces represents a value at a different location on a map. I would like to put two of them at a time side-by-side, and split them horizontally into 10368 separate 2x1 vectors, such that:

M = [ 1 3 5 7 9 ]

N = [ 0 2 4 6 8 ]

WANT:

MN1 = [ 1 0 ]

MN2 = [ 3 2 ]

MN3 = [ 5 4 ]

MN4 = [ 7 6 ]

MN5 = [ 9 8 ]

The end result should be a pair of different values at each location. Here's the kicker: since some of these vectors include NaN values, and I don't want to misplace any of the associated locations in the vectors, I would also like to return NaN values as placeholders for any incomplete pair of data, such that:

M = [ 1 3 NaN 7 9 ]

N = [ 0 2 4 6 NaN ]

WANT:

MN1 = [ 1 0 ]

MN2 = [ 3 2 ]

MN3 = NaN

MN4 = [ 7 6 ]

MN5 = NaN

How can I do this?

Answer 1

[Azzi Abdelmalek]

M = [ 1 3 NaN 7 9 ];
N = [ 0 2 4 6 NaN ];
idx=any(isnan([M;N]));
out=arrayfun(@(x,y) [x y],M,N,'un',0) ;
out(idx)={nan}

[Patrick]

That gives me a logical answer, not a vector answer; I need the actual values. When I used that, I got 10368 instances of "[1x2 double]", so it's close, but it seems like your code is simply telling me that the answers are 1x2 doubles, rather than giving me the values.

[Azzi Abdelmalek]

You can check the values

celldisp(out)

You should read about cell array

help cell

If you want the first element

out{1}

[Patrick]

Thank you, I understand now.

[Patrick]

How can I create a cell array without removing the NaN values?

[Azzi Abdelmalek]

What do you mean?

[Patrick]

Please see my new question. I have the cell array, but I need to perform a K-means analysis, and it is not working. I thought leaving the NaN values in might solve that problem, as the internal code for K-means in MATLAB automatically takes NaN values out of the equation, but I still have no luck.