Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

TypeError: iteration over a 0-d array Python

I am trying to write a very basic nearest neighbor calculation. I basically want to see what t looks like but I got this type error. When I asked the funciton to return just t it said "". When I asked it to turn to list it threw "TypeError: iteration over a 0-d array Python "

How do I fix this please?

...

t = np.array(map(lambda v:
             map(lambda w: distance(v, w, L), x_train.values),
             x_test.values)) 

...

Full trace: enter image description here

like image 739
Anna Huang Avatar asked Feb 06 '18 12:02

Anna Huang


2 Answers

It is possible to load a 0-d numpy array with arr1 = np.load(..., allow_pickle=True). To access the item stored as being a np.array, use:

arr1.item()

For instance, if the type stored is a dict d1 = {'item1':42, 'item2':np.array(1,3)} we can get the value as such: v1 = arr1.item()['item2'].

Important: Loading with allow_pickle=True is brings security risks and is not recommended.

like image 137
Tijs Maas Avatar answered Oct 31 '22 01:10

Tijs Maas


The problem is np.array does not take an iterator, you need convert to list first, as below:

t = np.array(list(map(lambda v: map(lambda w: distance(v, w, L),
                      x_train.values), x_test.values)))

As per numpy.array documentation, the required parameter must be:

An array, any object exposing the array interface, an object whose array method returns an array, or any (nested) sequence.

Alternatively, use numpy.fromiter and remember to supply dtype, e.g. dtype=float.

like image 43
jpp Avatar answered Oct 31 '22 00:10

jpp