Why use functools.wraps when writing decorators? | Interview Question

Q: Why use functools.wraps when writing decorators?

**Target Audience:** Developers who can write a basic decorator and want to know why the extra line matters **Difficulty:** Medium **Python version:** behaviour checked on CPython 3.11 (python3.11 and the in-page Pyodide 0.25.1 runtime); the 3.12 and 3.13 attribute list was printed on those interpreters > **How to read this doc:** Every idea is first explained in plain words. When a term has a precise name an interviewer expects to hear, it is introduced as a 📌 **Interview term** callout. The Quick Glossary at the end lists all of them. ## 1. Why This Even Matters — A Story First A hotel receptionist takes calls for every guest. If the receptionist answers the phone as "Reception" for everyone, callers reach the right person in the end, but the caller ID, the voicemail greeting and the phone directory all say "Reception". Nobody can tell who they actually reached. A decorator without `functools.wraps` is that receptionist. After `@log`, the name `area` points at the decorator inne

The problem

Question presented to candidate: "Most decorator examples put @functools.wraps(func) on the inner wrapper. What goes wrong if you leave it out, and what exactly does wraps do?"

What a strong answer should cover:

Without it, the decorated name points at the wrapper, so __name__, __qualname__, __doc__, __module__ and annotations describe the wrapper, not the original.
What breaks: logs and tracebacks that print names, help() and docs, inspect.signature, registries keyed by __name__, and pickle, which finds functions by module and qualified name.
What wraps does: it is functools.update_wrapper as a decorator; it copies WRAPPER_ASSIGNMENTS, merges WRAPPER_UPDATES (__dict__) and sets __wrapped__ to the original.
__wrapped__ lets inspect.signature report the real signature and lets inspect.unwrap or tests reach the undecorated function.
What it does not do: it does not change the wrapper code, defaults or behaviour; it only copies metadata. A version detail: 3.12 added __type_params__ to the copied attributes.

Code / implementation expected: Yes — a decorator with and without wraps, printing __name__, __doc__ and the signature.

decorator
wraps

Try it yourself

Sketch your approach before you read the solution. That is what the interview tests.

Stuck? Get a hint.

A short nudge toward the approach. It will not give the answer away.

Solution

Target Audience: Developers who can write a basic decorator and want to know why the extra line matters Difficulty: Medium Python version: behaviour checked on CPython 3.11 (python3.11 and the in-page Pyodide 0.25.1 runtime); the 3.12 and 3.13 attribute list was printed on those interpreters

How to read this doc: Every idea is first explained in plain words. When a term has a precise name an interviewer expects to hear, it is introduced as a 📌 Interview term callout. The Quick Glossary at the end lists all of them.

1. Why This Even Matters — A Story First

A hotel receptionist takes calls for every guest. If the receptionist answers the phone as "Reception" for everyone, callers reach the right person in the end, but the caller ID, the voicemail greeting and the phone directory all say "Reception". Nobody can tell who they actually reached.

A decorator withou

The solution is hidden so you can try first

10 min read · 9 sections

Run the code

Edit the code and run it here. Each console line with an expected value becomes a test.

PythonMetadata, signature and __wrapped__ with and without functools.wraps

Practice

PythonReal bugs: a name-keyed registry, pickling, and reaching the original

Practice

Discussion

Other approaches, follow-ups and how it went in real interviews.

Sign in to join the discussion.

No comments yet. Share how you would answer it.

The solution is hidden so you can try first

Keep your momentum going.