[FIXED] How to retrieve an object saved by numpy

Issue

I save an object of my class by numpy.save() and I can load it by numpy.load() but the value that I retrieve is: [<classes.SampleClass object at 0x7fb8ebd7a1f0>].

class SampleClass:
   
    def __init__(self, y):
        self.y = y


y=np.array([1,2,3])
x=SampleClass(y)
np.save("x",x, allow_pickle=True)

xx=np.load("x.npy")

This code is an example and xx value is [<classes.SampleClass object at 0x7fb8ebd7a1f0>]. How can I access to the values of y (xx.y)?

Solution

'allow_pickle' is required for the load, not the save.

In [2]: class SampleClass:
   ...: 
   ...:     def __init__(self, y):
   ...:         self.y = y
   ...: 
   ...: 
   ...: y=np.array([1,2,3])
   ...: x=SampleClass(y)
   ...: np.save("x",x)
   ...: xx=np.load("x.npy", allow_pickle=True)

In [2]: 

In [3]: xx
Out[3]: array(<__main__.SampleClass object at 0x000002583B8BCDC0>, dtype=object)

xx is a 0d object dtype array. It has one element, which can be extracted with

In [4]: xx.item()
Out[4]: <__main__.SampleClass at 0x2583b8bcdc0>

or:

In [5]: xx[()]
Out[5]: <__main__.SampleClass at 0x2583b8bcdc0>

The () index matches the 0d shape of the array.

Answered By - hpaulj