How to make NaN data an empty entry

I have a 2160 by 4320 data that has lots of NaN. How do I make matlab ignore the NaNs (i.e making them an empty entry, not zero) and still maintain the same size of the data size

Answer 1

[Walter Roberson]

A=randi(500,5,5);
A(A<100)=nan

A = 5×5

138 207 415 446 134 NaN 437 153 NaN 449 119 443 141 127 NaN 294 235 NaN 336 149 NaN 155 235 314 264

Ac = num2cell(A)

Ac = 5×5 cell array
    {[138]}    {[207]}    {[415]}    {[446]}    {[134]}
    {[NaN]}    {[437]}    {[153]}    {[NaN]}    {[449]}
    {[119]}    {[443]}    {[141]}    {[127]}    {[NaN]}
    {[294]}    {[235]}    {[NaN]}    {[336]}    {[149]}
    {[NaN]}    {[155]}    {[235]}    {[314]}    {[264]}

Ac(isnan(A)) = {[]}

Ac = 5×5 cell array
    {[     138]}    {[207]}    {[     415]}    {[     446]}    {[     134]}
    {0×0 double}    {[437]}    {[     153]}    {0×0 double}    {[     449]}
    {[     119]}    {[443]}    {[     141]}    {[     127]}    {0×0 double}
    {[     294]}    {[235]}    {0×0 double}    {[     336]}    {[     149]}
    {0×0 double}    {[155]}    {[     235]}    {[     314]}    {[     264]}

Ac is now the same size as A, but the nan entries have been made empty.

Answer 2

[Matt J]

It depends on the operation. If you are taking the maximum of each data colum, for example, the NaNs are already ignored, e.g.,

A=randi(500,5,5);
A(A<100)=nan

A = 5×5

156 367 474 104 134 395 166 128 NaN 108 478 360 196 392 387 160 150 NaN 439 190 336 NaN 219 148 330

maximum = max(A)

maximum = 1×5

478 367 474 439 387

[Opeyemi Kehinde]

I want to create a new variable of the same size and data but this new variable should show blanks/nothing in the places where nan occured. Something like find(isnan(var))=[]

[Matt J]

No, you can't do that. Some place holder to mark missing data will be present in any Matlab array type.

[Walter Roberson]

Well it can be done if it is intended for display only, by constructing a character array and disp() that.

[Opeyemi Kehinde]

Thanks for the help, but I don't think I can use it for a mathematical calculation (maybe multiplying with another matching data) in this cell format or plot a graph.

[Walter Roberson]

Suppose that somehow you were able to create a numeric-like matrix that you could do mathematics with, and yet displayed some elements as blanks/nothing. What mathematical properties should be blank/nothing have? For example should [blank*2,5] have the blank be treated as the empty vector, and empty times 2 is empty, and horzcat([], 5) is the scalar 5 ??

[Opeyemi Kehinde]

Ok thanks