What does print(h.cumsum(axis=1).tolist()) print? NumPy Reductions and the axis Argument

Answer, D. [[1, 3], [3, 7]]

A. [[1, 2], [4, 6]]

Accumulates down each column instead, the result for axis=0 on the same array.

B. [1, 3, 6, 10]

Flattens the array first, the result cumsum would give without any axis at all.

C. [[1, 2], [3, 4]]

Assumes cumsum leaves the array unchanged when an axis is given.

D. [[1, 3], [3, 7]]

Correct. axis=1 accumulates across each row, so the first row becomes 1 and 1+2, and the second row becomes 3 and 3+4.

cumsum(axis=1) resets its running total at the start of every row and walks left to right within that row only. Rows never influence each other under this axis, unlike axis=0 where columns accumulate independently of each other.