readmatrix produces NaN values after parsing a .txt file

Hi there guys,

I've encountered weird phenomena while working with readmatrix(). I produce several .txt files with emphemerides data in an orbital propagator software and want to work on the position and velocity data further in Matlab. The thing is that after parsing those .txt files with readmatrix() I receive weird NaNs in random places. For instance, for your convenience I will only provide matrixA with the least amount of NaNs, but there are other files where the number reaches up to 221, which after parsing results in 2 NaNs (see code section below).

Thing is, when I started to manually check my .txt files, I found out that there are no NaNs in txt files themselves. At the row/col id provided by the find(isnan()) routine, I expected to see smth weird in the txt file itself. However, what I can see is reasonable and expected number, so the problem is probably in the way how I set up readmatrix itself. Can you give me any suggestion on how to deal with this NaN generation issue? I will also attach an archived matrixA.txt for your convenience.

Thanks beforehand!

matrixA=readmatrix('matrixA.txt');
[row,col]=find(isnan(matrixA))

row =

       45636
       19918


col =

     3
     4
     

[darova]

I don't have readmatrix. I tried importdata

A = importdata('matrixA.txt');
data = A.data;
find( isnan(data(:)) )

result

ans =
   Empty matrix: 0-by-1

[h_chep]

Thank you for your solution! It works perfectly fine!

If you could edit this solution to an answer, I would gladly accept it as an answer.

[David Ramirez]

Thank you!

Answer 1

[darova]

See this solution

I don't have readmatrix. I tried importdata

A = importdata('matrixA.txt');
data = A.data;
find( isnan(data(:)) )

result

ans =
   Empty matrix: 0-by-1

Answer 2

[Ameer Hamza]

The data in your txt file is separated using an irregular number of spaces. For some reason, it confuses readmatrix in a few rows. Try this

data = readmatrix('matrixA.txt', ...
    'Delimiter', ' ', ...
    'ConsecutiveDelimitersRule', 'join');

[h_chep]

Thanks for your tip!

It works, but produces an extra column of NaNs somehow. But the solution from darova solved my issue.

[Ameer Hamza]

I didn't have this problem in R2020a. There might be some issues fixed in this release.

[Luis Fernando Gutierrez Urzua]

This solution worked for me in R2021b. Thanks!