Object casts convert to python type (which makes some operations work magically) #16876

seberg · 2020-07-15T16:02:25Z

Consider the NumPy float32:

f32 = np.float32(2)
type(f32.astype(object))  # gives `float`

which is fair enough in a sense. Our scalars are weird...

This also makes some stranger things work though:

class myf:
   def __radd__(self, other):
       return f"called backward with {type(other)}"

f32 + myf()
# "called backward with <class 'float'>"

This works by first dispatching to the ufunc machinery, which converts both to arrays, which then casts f32 to object and then calls the full operator (both ways in principle).

Now, that would enter a recursive call of course if it would not convert to the float32 to a python float, so that:

myf() + np.float128()

segfaults due to infinite recursion I guess. If you swap the order, things are more OK, although it probably crashes just as much if np.float128 was instead a subclass of float128...

The text was updated successfully, but these errors were encountered:

seberg · 2020-07-15T16:03:33Z

So there are actually two things here:

Our scalar binop logic is shady
Discuss whether converting float32 to float is actually correct for object casts? Although I guess that we are stuck with that...

seberg mentioned this issue Sep 9, 2020

object array creation new conversion to int #3804

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Object casts convert to python type (which makes some operations work magically) #16876

Object casts convert to python type (which makes some operations work magically) #16876

Uh oh!

Uh oh!

Object casts convert to python type (which makes some operations work magically) #16876

Object casts convert to python type (which makes some operations work magically) #16876

Comments

Uh oh!