Python dataclasses: frozen fields require an immutable value model

Operation contract

The receipt record uses a tenant string, integer identifier and tuple of integer amounts. That component choice avoids the shallow-freeze problem of a mutable list field. The fixture shows value equality, tuple aggregation and rejection of ordinary reassignment to its identifier.

Failure and ownership boundary

Frozen is not a complete security barrier, and a frozen dataclass containing a list would still expose list mutation. Annotations also do not validate that arbitrary caller values have the declared runtime types. Construct validated records at Python JSON validation: reject duplicate members and non-integer amounts rather than treating generated initialization as a schema validator.

Working program

from dataclasses import dataclass, FrozenInstanceError

@dataclass(frozen=True)
class ReceiptBatch:
    tenant: str
    identifier: int
    amounts: tuple[int, ...]

batch = ReceiptBatch("tenant-a", 41, (125, 75))
print(batch == ReceiptBatch("tenant-a", 41, (125, 75)))
print(sum(batch.amounts))
try:
    batch.identifier = 42
except FrozenInstanceError:
    print("field assignment rejected")

Output

True
200
field assignment rejected

Costs and limits

Generated equality compares components and can cost O(n) for an n-element tuple. Freezing references does not recursively copy input; this fixture uses immutable components already.

Common Mistakes

• Frozen fields do not freeze a nested list.
• Annotations do not enforce a runtime schema.

Connected lessons

Python tuples: immutable containers can still contain mutable state, Python JSON validation: reject duplicate members and non-integer amounts, Java records: value carriers and defensive copies.

Related Python operation checks

Python hash and equality: immutable dictionary keys, Python slots: attribute layout without deep immutability.

Follow the integrity boundary

Python dataclasses.replace: a new record can retain old mutable fields.