Describe the bug

When calling confusion_matrix with non-empty y_true and y_pred but an empty
sample_weight array, the function crashes during internal processing.

This appears to be caused by inconsistent handling of empty sample_weight
compared to all-zero sample_weight, which currently raises a ValueError.

Steps/Code to Reproduce

from sklearn.metrics import confusion_matrix

y_true = [0, 1]
y_pred = [0, 1]

confusion_matrix(y_true, y_pred, sample_weight=[])

Expected Results

Either
A clear ValueError indicating invalid sample_weight, OR
A documented and consistent behavior matching all-zero sample_weight.

In any case, the function should not crash internally.

Actual Results

The function crashes with an internal error related to array length mismatch
during confusion matrix construction.

Versions

System:
python: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
executable: C:\Users\HP\venv-sklearn\Scripts\python.exe
machine: Windows-11-10.0.26200-SP0

Python dependencies:
sklearn: 1.9.dev0
pip: 25.3
setuptools: 80.9.0
numpy: 2.4.0
scipy: 1.16.3
Cython: 3.2.4
pandas: None
matplotlib: None
joblib: 1.5.3
threadpoolctl: 3.6.0

Built with OpenMP: True