fast delete of rows in a table

Question

Francesco Giuseppe Fornari 2020-5-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/535278-fast-delete-of-rows-in-a-table

评论： Francesco Giuseppe Fornari 2020-5-27

采纳的回答： Image Analyst

在 MATLAB Online 中打开

hi

I've got a table with 120k rows, and I need to delete some rows if there is a specific condition.

I wrote this:

for i=1:(size(db,1)-1)
     if db.INTERVENTION(i)==1 && isequal(db.SETTLEMENTDATE(i-1),db.SETTLEMENTDATE(i)) && isequal(db.DUID(i-1),db.DUID(i))
            db(i-1,:)=[];
     end
end

This code takes a long time to execute.

Is there a faster way?

thanks!

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。

请先登录，再回答此问题。

Answer 1

Image Analyst 2020-5-27

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/535278-fast-delete-of-rows-in-a-table#answer_440253

在 MATLAB Online 中打开

Try this:

numRows = size(db,1)-1;
rowsToDelete = false(numRows, 1);
for k = 2 : numRows  % Has to start at 2, right?  Since you're using k-1.  Don't use i, the imaginary variable, as a loop index.
     if db.INTERVENTION(k)==1 && isequal(db.SETTLEMENTDATE(k-1),db.SETTLEMENTDATE(k)) && isequal(db.DUID(k-1),db.DUID(k))
            rowsToDelete(k) = true;
     end
end
db(rowsToDelete, :) = []; % Remove these rows.
% Or equivalently
%db = db(~rowsToDelete, :); % Extract everything BUT the rows to delete.

3 个评论
显示 1更早的评论隐藏 1更早的评论

Image Analyst 2020-5-27

I think the reason being is that when you removed a single row from the table, it had to rebuild the table each time. If you give it a vector, it has to rebuild the table only once.

Francesco Giuseppe Fornari 2020-5-27

yes, it makes a lot of sense ;)

请先登录，再进行评论。

Answer 2

darova 2020-5-27

1
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/535278-fast-delete-of-rows-in-a-table#answer_440108

在 MATLAB Online 中打开

Write indices and delete outside the for loop

ind = logical(db*0);
for i=1:(size(db,1)-1)
     if db.INTERVENTION(i)==1 && isequal(db.SETTLEMENTDATE(i-1),db.SETTLEMENTDATE(i)) && isequal(db.DUID(i-1),db.DUID(i))
            ind(i,:) = true;
     end
end
db(ind) = [];