Python numpy库的ndarray数据结构使用很方便,这里记录一下如何将其传递给C/C++代码,
直接上答案:
使用到的库:ctypes
C++代码部分(toPython.cpp):
#includeextern "C" void showNdarray(int* data, int rows, int cols) { for (int i = 0; i < rows; i++) { for (int j = 0; j < cols; j++) { printf("data[%d][%d] = %dn", i,j,data[i * rows + j]); } } }
将C++代码编译成动态链接库:
g++ -o topython.so -shared -fPIC toPython.cpp
Python代码部分:
import ctypes
import numpy as np
# 加载动态库
lcpp = ctypes.cdll.LoadLibrary
cpplib = lcpp("./topython.so")
def transfer_array_to_cpp() :
data = np.array([[1,2,3,4,5],
[2,4,6,8,0]])
dataptr = data.ctypes.data_as(ctypes.c_char_p)
rows, cols = data.shape
# 调用C++函数,将ndarray数据传递给C++
cpplib.showNdarray(dataptr, rows, cols)
if __name__ == '__main__' :
transfer_array_to_cpp()
Python代码放在和C++链接库同一个目录下,运行即可:
data[0][0] = 1 data[0][1] = 0 data[0][2] = 2 data[0][3] = 0 data[0][4] = 3 data[1][0] = 2 data[1][1] = 0 data[1][2] = 3 data[1][3] = 0 data[1][4] = 4



