Delete rows in struct based on number of elements in field

Question

DavidL88 2022-8-22

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1783540-delete-rows-in-struct-based-on-number-of-elements-in-field

评论： Rik 2022-8-22

I have a 1x68 struct with 7 fields - Scouts. I want to delete rows where the number of elements under a field 'Vertices' is less than 20. How do I do this? I tried;

Scouts.Vertices(numel(Scouts.Vertices<20)) = []

but got Error using < Too many input arguments.

I would prefer not to run a loop.

Also tried this but comes back as rows = []

[rows,~] = find(numel(Scouts.Vertices)>10);

2 个评论
显示无隐藏无

Stephen23 2022-8-22

"I have a 1x68 struct with 7 fields - Scouts. I want to delete rows ..."

Your structure only has one row.

DavidL88 2022-8-22

I mean the columns (1 - 68).

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Rik 2022-8-22

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1783540-delete-rows-in-struct-based-on-number-of-elements-in-field#answer_1030725

在 MATLAB Online 中打开

If you mean you want to remove elements from your struct, you're fairly close already. Scouts.Vertices returns a comma separated list. You need to feed that to a function that will capture the output.

You could do this:

tmp={Scouts.Vertices};
tmp=cellfun('prodofsize',tmp);
ind = find(tmp>10);

But using a loop would be much simpler.

You could also use structfun, but that would only hide the loop and increase complexity as well.

2 个评论
显示无隐藏无

Stephen23 2022-8-22

编辑：Stephen23 2022-8-22

在 MATLAB Online 中打开

"You could also use structfun, but that would only hide the loop..."

It would not even hide the loop: STRUCTFUN only operates on scalar structures:

S = struct('A',{1,2,3});
structfun(@sqrt,S)
Error using structfun
Inputs to STRUCTFUN must be scalar structures.

Rik 2022-8-22

I may have been thinking of arrayfun. I don't believe I have used either of them recently. I don't think I have ever used arrayfun, other than in proving it's slower than a loop. But thank you for the correction.

请先登录，再进行评论。