我需要以這種方式通過已知的形狀和引數創建 2D np 陣列：
import numpy as np
rows = 9
cols = 8
xrows = 3
xcols = 2
想要的結果：
([[ 1, 1, 4, 4, 7, 7, 10, 10],
[ 1, 1, 4, 4, 7, 7, 10, 10],
[ 1, 1, 4, 4, 7, 7, 10, 10],
[ 2, 2, 5, 5, 8, 8, 11, 11],
[ 2, 2, 5, 5, 8, 8, 11, 11],
[ 2, 2, 5, 5, 8, 8, 11, 11],
[ 3, 3, 6, 6, 9, 9, 12, 12],
[ 3, 3, 6, 6, 9, 9, 12, 12],
[ 3, 3, 6, 6, 9, 9, 12, 12]])
rows%xrows = 0
cols%xcols = 0
rows 是陣列的行數，cols 是陣列的列數，xrows 是每個切片的行數，xcols 是每個切片的列數。答案應該是通用的，不適用于此引數
uj5u.com熱心網友回復：
IIUC，您可以使用：
((np.arange(rows//xrows*cols//xcols, dtype=int) 1)
.reshape((rows//xrows,cols//xcols), order='F')
.repeat(xrows,0)
.repeat(xcols,1)
)
輸出：
array([[ 1, 1, 4, 4, 7, 7, 10, 10],
[ 1, 1, 4, 4, 7, 7, 10, 10],
[ 1, 1, 4, 4, 7, 7, 10, 10],
[ 2, 2, 5, 5, 8, 8, 11, 11],
[ 2, 2, 5, 5, 8, 8, 11, 11],
[ 2, 2, 5, 5, 8, 8, 11, 11],
[ 3, 3, 6, 6, 9, 9, 12, 12],
[ 3, 3, 6, 6, 9, 9, 12, 12],
[ 3, 3, 6, 6, 9, 9, 12, 12]])
轉載請註明出處，本文鏈接：https://www.uj5u.com/qita/461738.html
下一篇：基于串列組合的numpy乘法