Filter matrix rows based on value of function in Matlab

4 次查看（过去 30 天）
ARS
ARS 2013-2-18
评论： yair suari 2015-9-1
This seems like a simple question but I have been unable to find an answer anywhere. If I have a Matlab matrix "A" consisting of an arbitrary number of rows, how would I filter these rows based on the value of some function "f" (the argument of which is a row vector)? In other words, how would I keep only the rows of matrix A for which f is true? I tried
A(f(A(:)), :)
but to no success. Any help would be greatly appreciated.
So for example, if I have the matrix:
A =
1 1
1 2
1 3
1 4
2 1
2 2
2 3
2 4
3 1
3 2
3 3
3 4
4 1
4 2
4 3
4 4
And a function f that returns true only for inputs [1 2], [1 3], [2 4] and [3 4], my desired output would be:
ans =
1 2
1 3
2 4
3 4

请先登录，再回答此问题。

采纳的回答

Mark Whirdy
Mark Whirdy 2013-2-18
编辑：Mark Whirdy 2013-2-18
Shown below using an anonymous function for neatness, but the trick is just to make sure that your function (anonymous, nested, sub, or separate) returns a LOGICAL VECTOR (or a vector of linear indexes)
fn = @(x)(x>1);
a = randn(20,1);
fn(a); % this would return a logical vector if you ran it
myfilteredVector = a(fn(a));
If I've misunderstood, could you give me an example with an input & a required output.
------
So the bones of the required operation is actually a row-wise matching? I do this by transforming each row into a single unique element by 10*A(:,1) + A(:,2), as below. Then ismember() with its mex engine can then be put to work on matching individual element. There are other matching/indexing approaches though but the primary aim is to have the function outputting a logical vector.
A = [...
1, 1;
1, 2;
1, 3;
1, 4;
2, 1;
2, 2;
2, 3;
2, 4;
3, 1;
3, 2;
3, 3;
3, 4;
4, 1;
4, 2;
4, 3;
4, 4;];
fn = @(x)(ismember(x,[12;13;24;34]));
% fn(10*A(:,1) + A(:,2)) would return a logical vector
A(fn(10*A(:,1) + A(:,2)),:); % logical vector selects rows, & we want every column so (logic, :)
You can of course also specify the matchset a=[12;13;24;34] as an argument as below.
fn = @(x,a)(ismember(x,a));
a = [12;13;24;34];
A(fn(10*A(:,1) + A(:,2)),:);
To be honest, your function is kind of unnecessary really (unless there's a reason outside of the example you've given), as you can just use ismember with the unique encoder 10*A(:,1)+A(:,2).
a = [1,2;1,3;2,4;3,4];
A( ismember(10*A(:,1)+A(:,2), 10*a(:,1)+a(:,2)) , :)

更多回答（1 个）

Azzi Abdelmalek
Azzi Abdelmalek 2013-2-18
n=size(A,1);
A(arrayfun(@(x) expression(A(x,:)),(1:n)','un',0))=[]
  5 个评论
显示 3更早的评论
ARS
ARS 2013-2-18
Sorry if I'm misunderstanding, but I receive the error:
Error using subsindex
Function 'subsindex' is not defined for values of class 'cell'.
Azzi Abdelmalek
Azzi Abdelmalek 2013-2-18
编辑：Azzi Abdelmalek 2013-2-18
Try
n=size(A,1)
A(cell2mat(arrayfun(@(x) f(A(x,:)),(1:n)','un',0)),:)=[]
Also this is not a general solution, each expression has to be treated individually.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Logical 的更多信息

标签

产品

Community Treasure Hunt

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

Start Hunting!

Translated by