What is the difference between a list and a tuple, and when would you choose one over the other?
List vs tuple: mutability, semantics, and trade-offs
A list is a mutable sequence: you can add, remove, or replace elements in place. A tuple is an immutable sequence: once built, its element references cannot be changed. The deeper reason tuple exists is semantic and operational: immutability makes it hashable when all elements are hashable, so it can be used as a dict key or set member, and it signals a fixed record-like structure. CPython also stores tuples slightly more compactly because it does not need overallocation for growth.
- 1
Choose a list for homogeneous, growing/shrinking collections, queues/stacks at the end, or when in-place updates are expected.
- 2
Choose a tuple for fixed records, multiple return values, dict keys, set members, and read-only configuration.
- 3
Trade-off: namedtuple/dataclass can be clearer than a raw tuple for record semantics; dataclass gives mutability control and methods, namedtuple keeps tuple behavior.
- 4
Common mistake: thinking tuple immutability is deep. A tuple can contain a list, and that list can still be mutated.
- 5
Version note: dict/set key hashability depends on all tuple elements being hashable; this is not version-specific, but tuple memory/overallocation details are CPython implementation details.
Scenario Questions
0-2 years experience
- 1You need to return three values from a function and callers must not be able to modify the returned collection in place. What do you return and why?
- 2You are storing coordinates as dict keys. Which collection do you use and what restriction applies?
2-5 years experience
- 1A config loader returns a list of DB hosts. Another module accidentally appends to it and corrupts global config. How do you redesign the return type?
- 2You need to group records by a composite key of user_id and date. Which collection type for the key and why?
5-8 years experience
- 1You are designing an event schema with fixed fields but occasional optional metadata. Would you use tuple, namedtuple, dataclass, or dict? Discuss memory and API stability.
- 2A high-throughput pipeline emits millions of small records. How would list vs tuple affect memory and cache behavior?
8+ years experience
- 1Design an immutable value object that can be safely shared across threads and used as a cache key, while still supporting schema evolution.
- 2You need to expose a fixed-shape tensor-like result to C extensions. What Python sequence type minimizes overhead and why?
Follow-up Questions
- Can a tuple contain a list? What happens if you try to use that tuple as a dict key?
- When would you prefer a dataclass over a namedtuple or a plain tuple?