Removing NaNs from a struct

45 views (last 30 days)
Manny Kins
Manny Kins on 1 May 2019
Edited: Walter Roberson on 8 Apr 2020
I have a struct with the following layout:
T X Y
[0,1] [0, 1] [0,5]
[0] [2] [2]
NaN NaN NaN
NaN NaN NaN
NaN NaN NaN
NaN NaN NaN
[0] [3] [3]
Is there a way to remove all NaN values such that I am left with only:
T X Y
[0,1] [0, 1] [0,5]
[0] [2] [2]
[0] [3] [3]
I am using isnan but keep getting an error "Undefined function 'isnan' for input arguments of type 'struct'".
Thanks
  4 Comments
Manny Kins
Manny Kins on 1 May 2019
Just added, thanks
Walter Roberson
Walter Roberson on 1 May 2019
Probably the easiest way is to use struct2cell and cellfun(@(C) any(isnan(C(:)), thecell), and then any() that across the proper dimension, to arrive at a logical vector of location to remove. Then delete those from the original struct.

Sign in to comment.

Accepted Answer

Stephen23
Stephen23 on 1 May 2019
Edited: Stephen23 on 1 May 2019
This works for any sized fields (including empty), and gives you the choice of matching ALL or ANY of the fields containing NaN.
>> F = @(s)all(structfun(@(a)isscalar(a)&&isnan(a),s)); % or ANY
>> X = arrayfun(F,AllData.Passive);
>> AllData.Passive(X) = [];
checking:
>> AllData.Passive.T
ans =
0 1 2 3 4 6 7 8 9 10
ans =
0 3 4 7 9 10
ans =
0 2 6
ans =
0 1 6 7 8
ans =
10
ans =
10
ans =
10
ans =
10
  2 Comments
Manny Kins
Manny Kins on 2 May 2019
Thank you Stephen, this is really useful and flexible. I need to learn more about how and when to use the structfun and arrayfun functions as they seem to be very powerful
Yago Molano Gomez
Yago Molano Gomez on 7 Apr 2020
Hi! I've tried to copy this into my code but changing the empty brackets for a zero, and I get an error that says 'assignment between unlike types is not allowed'.
Also, when I use it with the empty brackets, it does not give me an error but does not remove the NaN's. How can I fix it? I'll attach the structures I'm referring to.

Sign in to comment.

More Answers (2)

Jos (10584)
Jos (10584) on 1 May 2019
TF = arrayfun(@(k) isnan(AllData.Passive(k).T(1)), 1:numel(AllData.Passive))
AllData.Passive(TF) = []
  1 Comment
Manny Kins
Manny Kins on 2 May 2019
Edited: Manny Kins on 2 May 2019
Thank you for your response Jos! Ahh Noddy (Noddy), the little man with the red and yellow car, good memories

Sign in to comment.


Felipe Ademir aleman hernandez
Edited: Walter Roberson on 8 Apr 2020
Hey, this works for me:
MyNewDataStruct = structfun( @rmmissing , MyDataStruct , 'UniformOutput' , false)

Categories

Find more on Creating and Concatenating Matrices in Help Center and File Exchange

Tags

Products


Release

R2019a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!