Search code examples
pythonpython-3.xctypes

Python ctypes dtype multiply

I have the following line in python:

a = (ctypes.c_float*10)()

Here a is initialized as an object with _repr_ result:

<__main__.c_float_Array_10 object at 0x7fbb3e6cf340>

My understanding is that this object is some sort of a pointer that can interoperate between python and C functions. How does this work and how are we able to multiply ctypes.c_float class with and integer?

Solution

  • c_float is a type corresponding to C float.
    c_float * 10 is an array of that type corresponding to C float[10].
    (c_float * 10)() is an instance of the 10-element float array.

    A DLL with a C function of float sum(float* arr, size_t size); could be called with:

    import ctypes as ct

dll = ct.CDLL('./somelibrary.dll')

# Inform ctypes of the type of arguments and return value.
dll.sum.argtypes = ct.POINTER(ct.c_float), ct.c_size_t
dll.sum.restype = ct.c_float

# make an instance of an array of 10 float values.
a = (ct.c_float * 10)(1.1, 2.2, 3.3, 4.4, 5.5, 6.6, 7.7, 8.8, 9.9, 10.0)
result = dll.sum(a, len(a))