Remove rows or cols whose elements are all NaN

Question

Jeon 2013-3-25

3
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/68510-remove-rows-or-cols-whose-elements-are-all-nan

评论： osman alper altun 2023-2-14

采纳的回答： Andrei Bobrov

在 MATLAB Online 中打开

How can I remove rows or cols whose elements are all NaN ? Withouot any dirty iterations?

For example,

A = [1 1 1 1 1 1 1 1 1 1;
     NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN;
     1 1 1 1 1 1 1 1 1 1;
     NaN NaN NaN NaN NaN NaN NaN NaN NaN NaN;];

should turned into

A = [1 1 1 1 1 1 1 1 1 1;
     1 1 1 1 1 1 1 1 1 1];

This is just an example. Actually I have a very big matrix. So I want a solution of this question to work well with my big matrix.

2 个评论
显示无隐藏无

Dev-iL 2014-7-13

编辑：Dev-iL 2014-7-13

在 MATLAB Online 中打开

Hi! You can find some clues here: A discussion of the opposite problem on SE. I personally achieved what you were attempting using:

A(~any(~isnan(A), 2),:)=[];

Walter Roberson 2015-9-20

Michal Gajewski commented

works

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Andrei Bobrov 2013-3-25

24
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/68510-remove-rows-or-cols-whose-elements-are-all-nan#answer_79794

编辑：Andrei Bobrov 2013-3-25

在 MATLAB Online 中打开

out = A(all(~isnan(A),2),:); % for nan - rows
out = A(:,all(~isnan(A)));   % for nan - columns

5 个评论
显示 3更早的评论隐藏 3更早的评论

Céldor 2015-4-2

在 MATLAB Online 中打开

Can I obtain a logical matrix I of indices to be held / removed something like

I = something here

and then use assign an altered matrices:

out1 = out1(I);
out2 = out2(I);
% ... etc.

Thanks

gringer45 2018-3-18

This doesn't really do what the question asks for. This selects all the columns or rows with none (zero) NaN values. So, this is answering the question: "Remove rows or cols whose elements have any (at least one) NaN"

请先登录，再进行评论。

Answer 2

Phillippe 2015-1-14

13
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/68510-remove-rows-or-cols-whose-elements-are-all-nan#answer_164900

To remove only ALL-NaN columns, do this instead:

A = A(:,~all(isnan(A)));

4 个评论
显示 2更早的评论隐藏 2更早的评论

Parth Dev Bundela 2022-10-22

Thanks man

osman alper altun 2023-2-14

Thank you!

请先登录，再进行评论。

Answer 3

Azzi Abdelmalek 2013-3-25

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/68510-remove-rows-or-cols-whose-elements-are-all-nan#answer_79793

在 MATLAB Online 中打开

A(isnan(A))=[]

5 个评论
显示 3更早的评论隐藏 3更早的评论

Serafeim Zacharopoulos 2020-5-27

To delete rows with all NaN's, maybe try this:

A = A(find(sum(~isnan(A)'))',:)

Walter Roberson 2020-5-28

在 MATLAB Online 中打开

A(all(isnan(A),2),:) = [];    %rows that are all nan
A(:, all(isnan(A),1)) = [];   %cols that are all nan

请先登录，再进行评论。

Answer 4

Saman 2013-10-23

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/68510-remove-rows-or-cols-whose-elements-are-all-nan#answer_100577

编辑：Saman 2013-10-23

在 MATLAB Online 中打开

Use this :

   out = A(:,any(~isnan(A)));  % for columns
   out = A(any(~isnan(A),2),:);   %for rows

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Answer 5

Alfaz Memon 2018-8-20

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/68510-remove-rows-or-cols-whose-elements-are-all-nan#answer_333368

编辑：Alfaz Memon 2018-8-21

在 MATLAB Online 中打开

input varibale : data

output variable : row_index( index of rows with all the value as NaN)

row_index( index of rows with all the value as NaN)

column_index( index of columns with all the value as NaN)

if(iscell(data))
x =find(cell2mat((cellfun(@(data) any(isnan(data),2),data,'UniformOutput',false)))); 
    [ia,ib] = ind2sub(size(data),x);  
    rows_unique = unique(ia);
    rows_unique(:,2)=histc(ia,rows_unique);
    row_index = rows_unique(find(rows_unique(:,2)==size(data,2)),1);
    columns_unique = unique(ib);
    columns_unique(:,2)=histc(ib,columns_unique);
    column_index = rows_unique(find(columns_unique(:,2)==size(data,1)),1);
else
   row_index  =find(~any(~isnan(data), 2)); % row with all NaN values
   column_index =find(~any(~isnan(data), 1)); %column with all NaN values 
end

2 个评论
显示无隐藏无

Walter Roberson 2018-8-20

Seems like a bit of a bother to just remove the rows or columns ?

I notice that you are using cellfun on the data, implying that the data is a cell array; in the original question it was a plain array.

Alfaz Memon 2018-8-21

编辑：Alfaz Memon 2018-8-21

yeah its for cell array. for plain array you can remove cellfun and just simply keep any(isnan(data),2) instead of x =find(cell2mat((cellfun(@(data) any(isnan(data),2),data,'UniformOutput',false))));

this can you work for cell array of different data type.

Also I have updated solution.

请先登录，再进行评论。