numpy array using python's long type

Question

I am trying to create a numpy array with type long (that is, Python's long, not numpy's long = int64).

If I do:

m = np.eye(6, dtype=long)
print(m.dtype)

This outputs int64; i.e., they are not Python longs.

Is there any way to create a numpy array with each element being of type long? Or is this some fixed-width versus non-fixed-width problem that numpy doesn't support? If so, is there any library (preferably with a nice C API like numpy's) that can do this?

Answer 1

Python's long integer type is not a native numpy type, so you will have to use the object data type. The elements of an numpy array with object type can be any python objects.

For example,

In [1]: x = np.array([1L, 2L, 3L], dtype=object)

In [2]: x
Out[2]: array([1L, 2L, 3L], dtype=object)

In [3]: x[0]
Out[3]: 1L

In [4]: type(x[0])
Out[4]: long

Whether or not this is useful for you depends on what you want to do with the array of longs.