int() vs .astype('int') Python

Question

I'm not sure when to use int() and when to use .astype('int'). Can anyone explain?

Is it just int() is used for single values and .astype('int') is used for vectors? I'm coming from an R background so I'm used to using as.integer

Answer 1

.astype() is a method within numpy.ndarray, as well as the Pandas Series class, so can be used to convert vectors, matrices and columns within a DataFrame. However, int() is a pure-Python function that can only be applied to scalar values.

For example, you can do int(3.14), but can't do (2.7).astype('int'), because Python native types don't have any such method. However, numpy.array([1.1, 2.2, 3.3]).astype('int') is valid.

(Strictly, it is also possible to define an __int__() method within one's own classes, which would allow int() to be applied to non-native types. Thanks to @juanpa.arrivillaga for pointing this out.)

Answer 2

astype is a numpy function, as @rwp points out. It is defined as:

def astype(self, typecode):
        ""
        return self._rc(self.array.astype(typecode))

._rc is defined as:

def _rc(self, a):
    if len(shape(a)) == 0:
        return a
    else:
        return self.__class__(a)

In English, this means that if your array has no shape -- i.e. it's a list -- it returns the array, else it returns the array itself cast to the indicated type.

int is a python builtin. It only deals with scalars.