NumPy: Row percentages with keepdims

Turn each row of counts into percentages of its own total, the classic use of keepdims, and check which column leads in each row.

What to build

A program that converts the rows of a table to percentages.

Requirements

  • Print totals (keepdims): shape (r, 1), then each row's percentages, 100 × value / row total, each to 1 decimal place, separated by spaces (nan for every value of a row whose total is 0).
  • Then print leaders: ..., the column with the largest percentage in each row (the first on ties), or - for a row whose total is 0.

Input

Line 1: r and c. Then r lines of c whole numbers, not negative.

Output

r + 2 lines.

Constraints

1 ≤ r, c ≤ 10; values ≤ 10^6.

Example

Input:

2 3
1 1 2
0 3 1

Output:

totals (keepdims): shape (2, 1)
25.0 25.0 50.0
0.0 75.0 25.0
leaders: 2 1

Why it works

Summing along axis 1 with keepdims gives an (r, 1) column that broadcasts across each row, so one division turns every row into fractions of its own total.

Examples

Example 1 · Input
2 3
1 1 2
0 3 1
Expected output
totals (keepdims): shape (2, 1)
25.0 25.0 50.0
0.0 75.0 25.0
leaders: 2 1
Example 2 · Input
2 2
0 0
5 5
Expected output
totals (keepdims): shape (2, 1)
nan nan
50.0 50.0
leaders: - 0

Prompts inside input() are ignored when checking, so input("n: ") is fine.

Hints

Hint 1

Each row's total is a column of shape (r, 1): divide every element of the row by its own total.

Hint 2

A row whose total is 0 gives nan for every element, as 0 / 0 does in NumPy.

Solution

Show one correct solution (try it yourself first)
r, c = map(int, input().split())
m = [[int(x) for x in input().split()] for _ in range(r)]
totals = [[sum(row)] for row in m]
print(f"totals (keepdims): shape ({len(totals)}, 1)")
leaders = []
for row, (t,) in zip(m, totals):
    if t == 0:
        print(" ".join("nan" for _ in row))
        leaders.append("-")
    else:
        pct = [100 * x / t for x in row]
        print(" ".join(f"{p:.1f}" for p in pct))
        leaders.append(str(pct.index(max(pct))))
print("leaders:", " ".join(leaders))

Lesson: Reductions along axes →