Remove rows or cols whose elements are all NaN

182 次查看（过去 30 天）
Jeon
Jeon 2013-3-25
评论： Andrew Sol 2024-11-14，7:43
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
Dev-iL
Dev-iL 2014-7-13
编辑：Dev-iL 2014-7-13
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),:)=[];

请先登录，再进行评论。

请先登录，再回答此问题。

采纳的回答

Andrei Bobrov
Andrei Bobrov 2013-3-25
编辑：Andrei Bobrov 2013-3-25
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
Céldor
Céldor 2015-4-2
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
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。

更多回答（6 个）

Phillippe
Phillippe 2015-1-14
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

请先登录，再进行评论。

Saman
Saman 2013-10-23
编辑：Saman 2013-10-23
Use this :
out = A(:,any(~isnan(A))); % for columns
out = A(any(~isnan(A),2),:); %for rows
Azzi Abdelmalek
Azzi Abdelmalek 2013-3-25
A(isnan(A))=[]
  6 个评论
显示 4更早的评论
Walter Roberson
Walter Roberson 2020-5-28
A(all(isnan(A),2),:) = []; %rows that are all nan
A(:, all(isnan(A),1)) = []; %cols that are all nan
Andrew Sol
Andrew Sol 2024-11-14，7:43
You proposed too simple a solution and you were wrong.

请先登录，再进行评论。

Alfaz Memon
Alfaz Memon 2018-8-20
编辑：Alfaz Memon 2018-8-21
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
Walter Roberson
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
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。

Ilham Hardy
Ilham Hardy 2013-3-25
Haven't tried this, but it should works:
A(isnan(A))=[];
  1 个评论
Jan
Jan 2013-3-25
This does not solve the wanted: "delete rows or cols whose elements are all NaN"

请先登录，再进行评论。

Nike
Nike 2013-3-25
Simplest is
A(isnan(A))= [];
  1 个评论
Jan
Jan 2013-3-25
This has been posted twice already. But it still does not solve the original question:
delete rows or cols whose elements are all NaN
For e.g. A = [1, NaN, 1; NaN, 1, NaN] nothing should be deleted.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Resizing and Reshaping Matrices 的更多信息

标签

Community Treasure Hunt

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

Start Hunting!

Translated by