Function with input an array of classes.

9 次查看（过去 30 天）
Eleftherios Ioannidis
Maybe it's because I have a lot of programming experience in C/C++, but I am trying to build a function inside a class that takes a whole array of classes as an input. Matlab always returns "too many arguments". This is how I tried to do it.
classdef agent
properties
r
v
id
end
methods
function y=mean_distance(a)
S=sum(norm(a.r-this.r));
y=S/(size(a)-1)
%I also need a random access iterator to a, like disp(a(i).r(1));
end
end
Thanks. EDIT: class should probably have a handle for this to work, but I am unfamiliar with handles. Reading some relative tutorials.

请先登录，再回答此问题。

采纳的回答

Andrew Newell
Andrew Newell 2011-3-18
I can't run your code because there are a lot of undefined terms, but I can see two problems. First, if you run
a(2).view(a(2),a)
it's like having 3 arguments. You should call your function using one of these two equivalent forms:
view(a(2),a)
a(2).view(a) % The a(2) is considered the first argument
Second,
n=size(agents)
will give you a vector [1 10], which will give you an error when you get to the loop
for i=1:n
...
Do you know how to use the debugger?

更多回答（4 个）

Andrew Newell
Andrew Newell 2011-3-18
To set values in this object, you must have a class constructor. Now suppose a is an array of length 2 with a.r having values 1 and 2. Then if you type
a.r
You'll get something like this:
ans =
1
ans =
2
It's a comma-separated list, not a vector. To bundle them together, you'll need square brackets, like this:
function y=mean_distance(a,this)
y=mean(norm([a.r]-[this.r]));
end
A handle is not necessary for this problem.
EDIT: By the way, I substituted mean for your method, but maybe you have some reason for dividing by length(a)-1 instead of length(a)?
Philip Borghesani
Philip Borghesani 2011-3-18
You have two common mistakes here:
  1. MATLAB class methods require an explicit this input
  2. obj.prop returns a comma separated list that must be concatenated to use as a vector input.
Fixed Example
classdef agent
properties
r
v
id
end
methods
function y=mean_distance(this,a)
S=sum(norm([a.r]-[this.r]));
y=S/(numel(a)-1);
%I also need a random access iterator to a, like disp(a(i).r(1));
end
end
end
A handle class is not needed for your example often simple small classes that are placed in vectors are better off as value classes.
Usage:
a=agent
a(1).r=5
a(2).r=5
b=a
b(2).r=10
mean_distance(a,b)
% or
a.mean_distance(b)
Eleftherios Ioannidis
This worked for this function. But in another function, I did everything you said (constructors, []brackets, explicit 'this') and I still get "too many arguments". Here is the function. The class properties are the same:
function agnts=view(this,agents)
%search for visible agents
constants;
n=size(agents)
xproduct1=0;
xproduct2=0;
agnts=[];
k=1;
for j=-J:dang:J
vseek=([cos(j) -sin(j);sin(j) cos(j)]*[this.v]')';
for i=1:n
xproduct2=det([vseek ; [agents(i).r]]);
if(xproduct2*xproduct1<0)
agnts(k)=i;
k=k+1;
end
xproduct1=xproduct2;
end
end
end
agents: an array of type agent(self class type) agnts: a uint8 vector. constants: everything not declared here is in constants.m.
Thanks for the immediate responce.
EDIT: I call it this way:
>>for i=1:10
a(i)=agent(i); %i==id
end
>>a(2).view(a(2),a);
Error: "too many input arguments"
%the same when I run: a(2).view(a); OR a(2).view([a]);
Eleftherios Ioannidis
Thanks Andrew, with your advice I managed to fix this. And Phillip too, the [] brackets saved me. It works fine now.

类别

在 Help Center 和 File Exchange 中查找有关 Data Type Identification 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