Numpy Zeros does not break zeros into array

Question

edit, originally i thought this was a problem with passing the class, but then realized it was because when i do val = np.zeros((x,y)) i can not assign a value to val[i] , i found the reason is because although the documentation http://docs.scipy.org/doc/numpy/reference/generated/numpy.zeros.html says >>>

np.zeros((2, 1))

returns

array([[ 0.],
       [ 0.]])

but when I do

ara = np.zeros((2,2))
print ara

i get

 [[ 0.  0.]
 [ 0.  0.]]

which is missing the ',' and makes it impossible to index

I have some class with some initialized values

class myclass():
    def __init__(self):
        self.stuffone = 0
        self.stufftwo = []

then i have some function

def myfunc(stuff):
    getstuff = myclass()
    getstuff.stuffone = stuff[0]
    getstuff.stufftwo = [stuff[1],stuff[0]]

    return getstuff

val = np.zeros((1,3))
val[0] = myfunc(stuff)

so I want to call myfunc, pass it some variables, then it should create a class instance and return it to me. What I see so far is that inside of myfunc, i print getstuff and it gives

<__main__.myclass instance at 0x000000000486CC08>

so that makes sense to me, however, when i return getstuff, the value that is returned is just junk, I suppose

[  1.09268349e-317   1.09268349e-317   1.09268349e-317]

yes sorry i was trying to make it simple and left that out. the code now is the full class, and in mkpp I took out the long parts of matrix stuff, and that is the full end. everything works in the mkpp function, and the print "MADE PP: " prints the instance, but once i pass it back its garbage — user-2147482637
– user-2147482637, Commented Dec 12, 2013 at 6:38
You were right in trying to make it simple (the full example is too complicated). The part where you went wrong is that you did not make the simplified example runnable and the larger example still has this same problem. — hugomg
– hugomg, Commented Dec 12, 2013 at 6:44
when i return getstuff, the value that is returned is just junk <-- please elaborate this, which is the actual problem. — Burhan Khalid
– Burhan Khalid, Commented Dec 12, 2013 at 6:48
actually as you see in the real code, it is spl[i], i tried it with just a variable and it is working. There is something wrong with the numpy zeros — user-2147482637
– user-2147482637, Commented Dec 12, 2013 at 6:58

Mailerdaimon · Accepted Answer · 2013-12-12 07:19:52Z

2

You can´t use a numpy array to store instances off classes like this. Look at the documentation!

=====
zeros
=====

Definition: zeros(shape, dtype=float, order='C')
Type: Function of numpy.core.multiarray module

[...]

Returns
-------
out : ndarray
    Array of zeros with the given shape, dtype, and order.

The array you create is of type float and not of type object.

Try:

val = numpy.zeros((1,3),numpy.dtype(object))
val[0] = myfunc(stuff)
print val[0,0]

EDIT

Pleae note!

This makes a difference:

>>> ara = np.zeros((2,2))
>>> ara
array([[ 0.,  0.],
       [ 0.,  0.]])
>>> print ara
[[ 0.  0.]
 [ 0.  0.]]

answered Dec 12, 2013 at 7:19

Mailerdaimon

6,1403 gold badges38 silver badges47 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

user4815162342 · Accepted Answer · 2013-12-12 07:20:54Z

2

There is nothing wrong with np.zeros(). You are missing the comma because you are using print ara, instead of just ara followed by enter. This makes Python print a "string" representation of the array, subtly different than the "repr" shown in the documentation. This display string has no effect on the behavior of the array.

Your code tries to put arbitrary "stuff" into a numpy array. Arrays don't work like that, they are typed and can only accept data of the proper type. np.zeros((2, 2)) returns a two-dimensional array of float64, so you can only place floats in it.

answered Dec 12, 2013 at 7:20

user4815162342

159k22 gold badges348 silver badges418 bronze badges

1 Comment

user-2147482637 Over a year ago

thank you! Mailerdaimon answered just before you but thats was the problem

Collectives™ on Stack Overflow

Numpy Zeros does not break zeros into array

2 Answers 2

Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related