Indexing and Slicing

Index Element

Four corner cells of a 2×3 grid accessed by row and column index. The replay walks a list of [i, j] coordinates, extracting i and j from each, then reads grid[i][j] to collect the four values.

By hand

With NumPy

np.array(grid) creates a 2-D array. The comma syntax a[i, j] reads the element at row i, column j directly — no intermediate row object needed. The snapshot shows the array then four individual cell reads.

naive.py 
grid = [[1, 2, 3], [4, 5, 6]]
coords = [[0, 0], [0, 2], [1, 0], [1, 2]]
cells = []
for coord in coords:
    i = coord[0]
    j = coord[1]
    cells.append(grid[i][j])
print('RESULT:', cells)
library.py 
import numpy as np

grid = [[1, 2, 3], [4, 5, 6]]
a = np.array(grid)
print('shape:', a.shape)
print('dtype:', a.dtype)
print('values:', a.tolist())
print('a[0, 0]:', int(a[0, 0]))
print('a[0, 2]:', int(a[0, 2]))
print('a[1, 0]:', int(a[1, 0]))
print('a[1, 2]:', int(a[1, 2]))
print('RESULT:', [int(a[0, 0]), int(a[0, 2]), int(a[1, 0]), int(a[1, 2])])

shape: (2, 3)
dtype: int64
values: [[1, 2, 3], [4, 5, 6]]
a[0, 0]: 1
a[0, 2]: 3
a[1, 0]: 4
a[1, 2]: 6
RESULT: [1, 3, 4, 6]

Implementation notes

  • Python's a[i][j] first retrieves the row a[i] (a full Python list), then indexes into it with [j]. NumPy's a[i, j] is a single operation that computes the flat offset i * cols + j and reads one element from the contiguous buffer — the same arithmetic as reshape-1d-2d's r * cols + c.
  • Shape, dtype, and values are shown explicitly here because ndarray.__repr__ output varies with NumPy version and print options.