You can do this using the datasample function:

http://www.mathworks.it/it/help/stats/datasample.html

For example, to extract three values from the vector A:

y = datasample(A, 3, 'Replace', false)

This works also if A is a matrix, an N-dimensional array or a dataset array.

If for any reason you have not access to this function, you can use randperm:

y = A(randperm(length(A), 3))

In this case, if A is a matrix and not a vector, you have to work on a vectorized version of it: A = A(:).

Hope this helps.

Another possibility:

y = A(randi(numel(A),3,1))

Datasample and randsample can be used.

Dear Simone Scardapane, Mattia Saccoccio & Deepak Bharti

Yes this is a perfect solution.

I have solved by using this

n2=[4 6 9];

f2=[0 1 1 0 1]

n11=([n2(find(f2==0))]);

Thank You very much

Best answer is

>> n = 3;

>> A = [1 2 3 4 5];

>> randsample(A, n)

Does anyone know if randsample can be nested safely with parfor?  I seem to recal a problem with random number generation between parfor processes.