python函数返回数组_如何使用ctypes将C函数中返回的2d数组传递给python

我有一个C函数返回2d数组(实际上是double **),我想使用ctypes将此double **传递给python.对此有疑问,但是在参数列表中而不是使用函数的返回值来修改数组.这是我的代码(C函数和相应的python包装器)

double** flux_function_2d(double *u,double gamma){

double **F;

double p,H;

F &＃61; (double **)malloc(4*sizeof(double *));

for (int i &＃61; 0; i <4; i&＃43;&＃43;)

F[i] &＃61; (double *)malloc(2*sizeof(double));

p &＃61; (gamma - 1.0) * (u[3] - 0.5 * (u[1]*u[1] &＃43; u[2]*u[2]) / u[0]);

H &＃61; u[3] / u[0] &＃43; p / u[0];

F[0][0] &＃61; u[1]; F[1][0] &＃61; u[1] * u[1] / u[0] &＃43; p;

F[2][0] &＃61; u[1] * u[2] / u[0]; F[3][0] &＃61; u[1] * H;

F[0][1] &＃61; u[2]; F[1][1] &＃61; u[2] * u[1] / u[0];

F[2][1] &＃61; u[2] * u[2] / u[0] &＃43; p; F[3][1] &＃61; u[2] * H;

return F;

}

def c_flux_function_2d(u,gamma):

flux &＃61; np.ctypeslib.load_library("libs/flux.so",".")

flux.flux_function_2d.argtypes &＃61; [POINTER(c_double),c_double]

flux.flux_function_2d.restype &＃61; POINTER(POINTER(c_double))

F &＃61; flux.flux_function_2d(u.ctypes.data_as(POINTER(c_double)),c_double(gamma))

F_arr &＃61; np.ctypeslib.as_array(F,shape&＃61;(4,2))

return F_arr

该错误发生在F_arr &＃61; np.ctypeslib.as_array(F,shape &＃61;(4,2))处,这意味着numpy无法解析ctypes中的指针的指针.

提前致谢&＃xff01;