TypeError: unsupported operand type

Python TypeError "unsupported operand type" has a confusing variant in NumPy and Pandas code. NumPy arrays and Pandas Series silently coerce types in ways that differ from native Python. Adding a Python integer to a NumPy int8 array may silently overflow, while mixing NumPy types with Python types raises TypeError in certain NumPy 2.0 operations where NumPy 1.x was lenient. NumPy 2.0 (released 2024) introduced stricter type casting than 1.x. Many data science scripts upgraded from NumPy 1.24 to 2.0 break with TypeError in arithmetic operations that previously worked. The key change: implicit downcast from float64 to int32 (and similar) is now rejected. Check your NumPy version with import numpy; print(numpy.__version__) before debugging. A second edge case: dividing by None raises TypeError rather than the more intuitive ZeroDivisionError. This appears when DataFrame rows with missing values interact with boolean logic that returns Python None rather than numpy.nan (which is a float). For Pandas operations, df.astype({'column': 'float64'}) before arithmetic ensures type consistency. When mixing Python native types with NumPy types, int(np_value), float(np_value), or np_value.item() converts NumPy scalar types to native Python types, avoiding cross-type operation errors in NumPy 2.0.