How does exception handling work in Python?

Question

Accepted Answer

Python handles errors with **`try`/`except`** blocks, plus optional `else` and `finally`. Python embraces the **EAFP** principle — "Easier to Ask Forgiveness than Permission" — preferring to try an operation and handle exceptions over checking conditions first.

## Basic structure

```python
try:
    result = risky_operation()
except ValueError as e:           # catch a SPECIFIC exception type
    print(f"bad value: {e}")
except (KeyError, IndexError):    # catch multiple types
    print("lookup failed")
except Exception as e:            # catch-all (use sparingly, last)
    print(f"unexpected: {e}")
else:
    print("ran only if NO exception occurred")
finally:
    cleanup()                     # ALWAYS runs (success or failure)
```

- **`except SpecificError`** — catch specific types (preferred over a bare catch-all).
- **`else`** — runs only if no exception was raised.
- **`finally`** — always runs (cleanup, releasing resources), even if an exception propagates.

## EAFP vs LBYL — the Pythonic style

```python
# LBYL ("Look Before You Leap") — check first
if key in my_dict:
    value = my_dict[key]

# EAFP ("Easier to Ask Forgiveness") — try it, handle failure — more Pythonic
try:
    value = my_dict[key]
except KeyError:
    value = default
```

Python idiomatically favors **EAFP** — attempt the operation and catch the exception — which avoids race conditions and is often cleaner than pre-checking.

## Raising and custom exceptions

```python
if amount < 0:
    raise ValueError("amount must be positive")    # raise a built-in

class InsufficientFundsError(Exception):           # custom exception type
    pass
raise InsufficientFundsError("balance too low")
```

## Best practices

```python
# ❌ bare except hides bugs (catches EVERYTHING, even typos/KeyboardInterrupt)
try: ...
except: pass

# ✅ catch specific exceptions; re-raise what you can't handle
try: ...
except ValueError as e:
    logger.error(e)
    raise                  # re-raise to propagate
```

```text
✓ Catch SPECIFIC exceptions, not bare except
✓ Don't silently swallow errors (except: pass) — log or handle them
✓ Use finally / context managers (with) for cleanup
✓ Raise meaningful, specific exception types
```

## Why it matters

Robust error handling is essential for reliable programs, and Python's approach has distinct idioms worth knowing: the `try/except/else/finally` structure, the **EAFP** philosophy (try-and-handle rather than check-first, which is more Pythonic and avoids race conditions), catching *specific* exceptions (not bare `except`, which hides bugs), and using `finally`/context managers for cleanup.

Understanding these prevents both crashes from unhandled errors and the subtle, hard-to-debug problems caused by over-broad or silent exception handling — a frequent source of unreliable code.