BUG: Fix issue #61841 - .rolling().mean() returns NaNs on reassignment

[abujabarmubarak]

This pull request fixes issue #61841, where reassigning a .rolling().mean() result unexpectedly returns a Series of all NaNs, even after copying the DataFrame.

---

🐛 Bug Reproduction Example

import pandas as pd
import numpy as np

df = pd.DataFrame({"Close": np.arange(1, 31)})
df = df.copy()
df["SMA20"] = df["Close"].rolling(20).mean()
df["SMA20"] = df["Close"].rolling(20).mean()  # ❌ Returns NaNs

---

🔧 What Was Fixed

• Modified logic in Window._apply():

  • Previously, result slicing ([:: self.step]) broke shape/index alignment.
  • Now it checks for self.step and slices only after full shape result is returned.

# ✅ Fixed
result = self._apply_columnwise(...)
if self.step is not None and self.step > 1:
    if isinstance(result, Series):
        result = result.iloc[:: self.step]
    elif isinstance(result, DataFrame):
        result = result.iloc[:: self.step, :]
return result

• Moved Series and DataFrame imports to the top level of rolling.py to fix pre-commit check failures related to inconsistent namespace usage.

---

🧪 How Verified

import pandas as pd
import numpy as np

df = pd.DataFrame({"Close": np.arange(1, 31)})
df = df.copy()
df["SMA20"] = df["Close"].rolling(20).mean()
df["SMA20"] = df["Close"].rolling(20).mean()
print(df.tail())  # ✅ Correct output

---

✅ Status

• Bug fixed
• Code passes all CI and pre-commit checks
• Imports are consistently handled

Thanks for reviewing this PR!