Is it possible to copy the elements of a matrix to a set of variable names ?

Given is a matrix A of size [1x6]. What I want is to copy the elements of A to a set of variable names [a,b,c,d,e,f], so that

a = A(1);
b= A(2)
a.s.o.

I tried

[a,b,c,d,e,f]  = deal(A);

But this copies not the elements but the hole matrix A into each variable.

Any help is much appreciated.

Answer 1

[Walter Roberson]

Acell = num2cell(A);
[a,b,c,d,e,f]  = deal(Acell{:});    %older MATLAB
[a,b,c,d,e,f]  = Acell{:};          %newer MATLAB

[Alexander]

Thanks Walter. This is quite simple. And I do not have any problems with definition of what workspace it is.

Thank you, Azzi for your help too.

Answer 2

[Azzi Abdelmalek]

A=rand(1,6)
s='abcdef'
for k=1:numel(s)
 assignin('base',s(k),A(k))
end

[Alexander]

Thanks Azzi.

What I ment was, is there a single function, as for instance deal, to avoid a for-loop?

[Azzi Abdelmalek]

A=rand(1,6)
s='abcdef'
arrayfun(@(x) assignin('base',s(x),A(x)),1:numel(s))

[Azzi Abdelmalek]

%or

arrayfun(@(x,y) assignin('base',x,y),'a':'f',A)

[Alexander]

Ok, this work fine. Thanks.

Now I have one additional problem with the workspace definition. I have put this code into a function. And neighter 'base' nor 'caller' produces the variables in the function workspace but in the base workspace. So calling the function produces an error.

[Azzi Abdelmalek]

Post your function

[Alexander]

% generates function handle to bimodal gaussian distribution:
% f(X; A1, A2, x01, x02, sig1, sig2) = ...
%  A1*exp( -(X-x01).^2 / (2*sig1^2) )  +  A2*exp( -(X-x02).^2 / (2*sig2^2) )
% beta is the coefficient vector [A1,A2,x01,x02,sig1,sig2]

function f = bimodgauss(beta)

s={'A1','A2','x01','x02','sig1','sig2'};
for k=1:numel(s)
assignin('caller',s{k},beta(k))
end

f = @(X) A1*exp(-(X-x01).^2 / (2*sig1^2))  +  A2*exp(-(X-x02).^2 / (2*sig2^2));