Remove rows or cols whose elements are all NaN
조회 수: 182 (최근 30일)
이전 댓글 표시
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.
댓글 수: 2
채택된 답변
Andrei Bobrov
2013년 3월 25일
편집: Andrei Bobrov
2013년 3월 25일
out = A(all(~isnan(A),2),:); % for nan - rows
out = A(:,all(~isnan(A))); % for nan - columns
댓글 수: 5
Céldor
2015년 4월 2일
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
2018년 3월 18일
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"
추가 답변 (6개)
Phillippe
2015년 1월 14일
To remove only ALL-NaN columns, do this instead:
A = A(:,~all(isnan(A)));
댓글 수: 4
Azzi Abdelmalek
2013년 3월 25일
A(isnan(A))=[]
댓글 수: 6
Walter Roberson
2020년 5월 28일
A(all(isnan(A),2),:) = []; %rows that are all nan
A(:, all(isnan(A),1)) = []; %cols that are all nan
Alfaz Memon
2018년 8월 20일
편집: Alfaz Memon
2018년 8월 21일
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
댓글 수: 2
Walter Roberson
2018년 8월 20일
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
2018년 8월 21일
편집: Alfaz Memon
2018년 8월 21일
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.
참고 항목
카테고리
Help Center 및 File Exchange에서 Resizing and Reshaping Matrices에 대해 자세히 알아보기
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!