pandas: A rolling mean with missing values

Model Series.rolling(k).mean(), the moving average used to smooth time series, including how it treats missing values.

What to build

A program that prints a rolling mean.

Requirements

  • For each position, the result is the mean of the window of the last k values ending there, printed with two decimals.
  • With the default min_periods (equal to k), the result is NaN when the window has fewer than k values (the first k - 1 positions) or holds a missing value. With min_periods M given, the mean of the window's present values is used when at least M are present, and NaN otherwise.
  • Print LABEL: RESULT for each position.

Input

Line 1: k and optionally M. Line 2: n. Then n lines LABEL VALUE, where VALUE is a number or NaN.

Output

n lines.

Constraints

1 ≤ k ≤ n ≤ 100; 1 ≤ M ≤ k.

Example

Input:

3
5
d1 1
d2 2
d3 3
d4 4
d5 5

Output:

d1: NaN
d2: NaN
d3: 2.00
d4: 3.00
d5: 4.00

Why it works

A rolling window looks back k positions from each row, and min_periods decides how many real values a window needs, which is how pandas keeps a few gaps from wiping out a whole moving average.

Examples

Example 1 · Input
3
5
d1 1
d2 2
d3 3
d4 4
d5 5
Expected output
d1: NaN
d2: NaN
d3: 2.00
d4: 3.00
d5: 4.00
Example 2 · Input
2
4
a 10
b NaN
c 30
d 40
Expected output
a: NaN
b: NaN
c: NaN
d: 35.00

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

Hints

Hint 1

For each position i, the window is the values from i - k + 1 to i.

Hint 2

By default a window with fewer than k values, or with a missing value, gives NaN (min_periods equals the window size).

Solution

Show one correct solution (try it yourself first)
parts = input().split()
k = int(parts[0])
min_periods = int(parts[1]) if len(parts) > 1 else k
n = int(input())
labels = []
values = []
for _ in range(n):
    label, value = input().split()
    labels.append(label)
    values.append(None if value == "NaN" else float(value))
for i in range(n):
    window = values[max(0, i - k + 1) : i + 1]
    present = [v for v in window if v is not None]
    if len(present) >= min_periods:
        print(f"{labels[i]}: {sum(present) / len(present):.2f}")
    else:
        print(f"{labels[i]}: NaN")

Lesson: Series: one labelled column →