Creating numpy array from c-pointer crashes ipython kernel

Question

With this code I am trying to create a numpy array from a malloc'ed c pointer, inspired by a blog post from Gaël Varoquaux.

The line that creates the array seems to lead to an invalid memory access as it crashes the kernel. What did I do wrong?

%%cython -f

from libc.stdlib cimport malloc
import numpy as np
cimport numpy as np

cdef array_from_pointer(double* ptr, int size):
    cdef np.npy_intp shape_c[1]
    shape_c[0] = <np.npy_intp> size
    ndarray = np.PyArray_SimpleNewFromData(1, shape_c, np.NPY_FLOAT64, ptr)
    return ndarray


cdef N = 12
cdef double* ptr = <double*> malloc(sizeof(double)*N)

array_from_pointer(ptr, N)

related: https://stackoverflow.com/a/25496494/2272172 – cel Mar 11 '20 at 18:08 — cel, Mar 11 '20 at 18:08

score 0 · Accepted Answer · answered Mar 11 '20 at 18:08

Before using numpy's Array API, you need to call np.import_array():

%%cython -f

from libc.stdlib cimport malloc
import numpy as np
cimport numpy as np
np.import_array()

cdef array_from_pointer(double* ptr, int size):
    cdef np.npy_intp shape_c[1]
    shape_c[0] = <np.npy_intp> size
    ndarray = np.PyArray_SimpleNewFromData(1, shape_c, np.NPY_FLOAT64, ptr)
    return ndarray


cdef N = 12
cdef double* ptr = <double*> malloc(sizeof(double)*N)

array_from_pointer(ptr, N)

Failing to do so causes the segmentation fault that crashes the ipython kernel.

Creating numpy array from c-pointer crashes ipython kernel

1 Answers1