[Numpy-discussion] How do I make a diagonal matrix?

Alan G Isaac aisaac at american.edu
Fri Jun 23 10:50:13 CDT 2006


On Fri, 23 Jun 2006, Sven Schreiber apparently wrote: 
>>>> help(n.diag) 
> Help on function diag in module numpy.lib.twodim_base: 
> diag(v, k=0) 
>     returns the k-th diagonal if v is a array or returns a array 
>     with v as the k-th diagonal if v is a vector. 

That is pretty damn obscure.
Apparently Travis's new doc string did not survive?
The Numpy book says:
diag (v, k=0)
    Return the kth diagonal if v is a 2-d array, or returns 
    an array with v as the kth diagonal if v is a 1-d array.

That is better but not great.
I think what is wanted is:
diag (v, k=0)
    If v is a 2-d array:
    return a copy of the kth diagonal of v (as a 1-d array).
    If v is a 1-d array:
    return a 2-d array with a copy of v as the kth diagonal 
    (and zeros elsewhere).

fwiw,
Alan Isaac

PS As a response to the question, it might be worth noting 
the following.
>>> y=N.zeros((5,5))
>>> values=N.arange(1,6)
>>> indices=slice(0,25,6)
>>> y.flat[indices]=values
>>> y
array([[1, 0, 0, 0, 0],
       [0, 2, 0, 0, 0],
       [0, 0, 3, 0, 0],
       [0, 0, 0, 4, 0],
       [0, 0, 0, 0, 5]])

Generalizing we end up with the following (from pyGAUSS):

def diagrv(x,v,copy=True):
	if copy: x = numpy.matrix( x, copy=True )
	else: x = numpy.matrix( x, copy=False )
	stride = 1 + x.shape[1]
	x.flat[ slice(0,x.size,stride) ] = v
        return x






More information about the Numpy-discussion mailing list