How to create n arrays from a mxn matrix using mat2dataset

Hi guys,

I'm using mat2dataset on a 100x100 matrix and i want to create 100 arrays of each column of the dataset. This is what i'm doing:

% y=100x100 matrix  
s=mat2dataset(y);
  y1=(ds.y1);
  y1=y1.';
  y2=(ds.y2);
  y2=y2.'; ... %and so on

There must be a way to do this avoiding writing every single array!

Hope you can help me

Thanks!

Answer 1

[Steven Lord]

DON'T DO THIS.

Seriously, DO NOT DO THIS.

See question 1 in the Programming section of the FAQ for more information about why this is a Bad Idea. Dynamic field names (described in the FAQ question) work for dataset arrays just like they do for struct arrays, if I remember correctly.

[Gaspar Cid]

Al'right, i knew that was a bad idea doing this, thank you. I read the FAQ and i guess i just should use the dataset. The problem is that i want to evaluate each value of y (100 arrays) on this function

[u v w] = yang(x0,y0,z0,a,A,P_G,mu,nu,theta,phi,x,y,z)

How can i do it without writing 100 arrays? I need the 100 u v w solutions for the 100 "y"

Thanks

[Steven Lord]

load hospital % A MAT-file containing a dataset 
variableNames = hospital.Properties.VarNames;
agesAsDouble = hospital.(variableNames{3}); % Age is variable #3
agesAsDataset = hospital(:, 3); % New dataset containing only the ages

Create u, v, and w as arrays or cell arrays (depending on how homogeneous the outputs from your calls are going to be in terms of size and type) and use one of the techniques I used above to create agesAsDouble or agesAsDataset to iterate, storing the output for each variable in y in the corresponding element of u, v, and w.

[Gaspar Cid]

Thanks for your answer Steven.

This means that i have to define every y (from 1 to 100) as this?

y1=ds.(variableNames{1};
y2=ds.(variableNames{2},
...
y100=ds.(variableNames{3};

I think i'm not getting it.

I need the function yang to read every column of the dataset (100x100) as an array (only for y), because the other inputs are already 1x100 doubles.

Something like this

[u v w] = yang(x0,y0,z0,a,A,P_G,mu,nu,theta,phi,x,(ds.(variableNames{1-100}),z)

I tried using arrayfun or structfun, but there's always an error...

This is the entire code i'm using

clear all; close all; clc; 
scrsz = get(0,'ScreenSize');

x0 = 3774; y0 = -352; z0 = 6224;                                                  
a = 6204;                                                                   
P_G = 3.59e-3;                                                                      
mu = 3.2E10;                                                                 
nu = 0.282756;                                                                    theta = 26; phi = 109;                                                        

N=20000;
n=100;
x=linspace(-N,N,n);
y=[];
for i=1:n;
PHI(i)=(360./n).*i;
m(i)=tand(PHI(i));
c(i)= y0-m(i)*x0;
y(i,:)=m(i)*x+c(i);
plot(x,y(i,:))
hold on
end 

y=y.';
ds=mat2dataset(y);

x0=repmat(x0,1,(length(ds)));
y0=repmat(y0,1,(length(ds)));
z0=repmat(z0,1,(length(ds)));
a=repmat(a,1,(length(ds)));
P_G=repmat(P_G,1,(length(ds)));
mu=repmat(mu,1,(length(ds)));
nu=repmat(nu,1,(length(ds)));
theta=repmat(theta,1,(length(ds)));
phi=repmat(phi,1,(length(ds)));
z=repmat(0,1,(length(ds))); 

A = 0.06;
A=repmat(A,1,(length(ds)));% geometric aspect ratio

[u v w] = yang(x0,y0,z0,a,A,P_G,mu,nu,theta,phi,x,y,z);

[Steven Lord]

You don't need to use a dataset here. Just use a for loop iterating over the columns of y, and store the results of each iteration through the for loop in an element of the cell arrays u, v, and w.

y = randi(10, 1, 20);
C = cell(1, 20);
for k = 1:20
  C{k} = magic(y(k));
end

C is a cell array whose nth cell contains the magic square matrix of size y(n) [or a specific 2-by-2 matrix if y(n) is 2, even though that matrix doesn't satisfy the requirements to be a magic square.]

[Gaspar Cid]

Al'right! I finally did this:

u = cell(1,200);
v= cell (1,200)
w=cell(1,200)
for k=1:200
[u{k} v{k} w{k}] = yang(x0,y0,z0,a,A,P_G,mu,nu,theta,phi,x,y(k),z)
end

And it seems to work! Thank you very much!