@dataclass parameters

Quick reference for the decorator parameters on @dataclass and the arguments to dataclasses.field(). For the tutorial treatment see the data classes notebook.

Decorator parameters

@dataclass(
    init=True,          # generate __init__
    repr=True,          # generate __repr__
    eq=True,            # generate __eq__
    order=False,        # generate __lt__, __le__, __gt__, __ge__
    unsafe_hash=False,  # generate __hash__ even when it would be risky
    frozen=False,       # make instances immutable
    match_args=True,    # (3.10+) expose __match_args__ for match/case
    kw_only=False,      # (3.10+) all fields are keyword-only
    slots=False,        # (3.10+) use __slots__
    weakref_slot=False, # (3.11+) add a __weakref__ slot
)
class Thing:
    ...

Parameter	Default	Effect
init	True	Generate __init__. Set False if you want to write your own.
repr	True	Generate __repr__.
eq	True	Generate __eq__ that compares all fields in order.
order	False	Generate __lt__/__le__/__gt__/__ge__. Requires eq=True.
unsafe_hash	False	Force a generated __hash__ even when the combination of eq/frozen would normally leave hashing off. Rarely needed.
frozen	False	Make instances immutable. Assigning to a field raises FrozenInstanceError. Frozen classes get __hash__ automatically.
match_args	True	Generate __match_args__ (Python 3.10+) so the class can be used in match patterns.
kw_only	False	Make every field keyword-only. Useful when you have a lot of fields and want callers to name them at call sites.
slots	False	Generate a __slots__ class. Smaller instances, rejects undeclared attributes. Note: this replaces the class with a new one, so some metaclass or weakref tricks need extra care.
weakref_slot	False	With slots=True, reserve a slot for weak references.

field() — per-attribute options

from dataclasses import dataclass, field

@dataclass
class Item:
    name: str
    tags: list[str] = field(default_factory=list)
    id_: int = field(default=0, repr=False)
    _cache: dict = field(default_factory=dict, compare=False, init=False)

Argument	Purpose
default=	A simple default value (immutable only).
default_factory=	A zero-arg callable called fresh for each instance. Use for list, dict, set, or any mutable default.
init=	If False, the field is not a parameter to __init__. Usually combined with default or default_factory.
repr=	If False, the field is omitted from the generated __repr__. Useful for noisy or sensitive fields.
compare=	If False, the field is ignored by __eq__ and the ordering methods.
hash=	If False, the field is omitted from __hash__. Usually you want compare and hash aligned.
metadata=	A read-only mapping for third-party tools. Dataclass itself doesn't consult it.
kw_only=	(3.10+) Per-field keyword-only control, overriding the class-level kw_only.

Related utilities

From the dataclasses module:

Function	Purpose
fields(cls_or_instance)	Returns the tuple of Field objects. Iterable, introspectable.
asdict(instance)	Recursively converts to a dict (nested dataclasses too).
astuple(instance)	Recursively converts to a tuple.
replace(instance, **changes)	Returns a new instance with some fields changed. Essential for frozen=True classes.
make_dataclass(name, fields, ...)	Creates a dataclass dynamically. Rarely needed.
is_dataclass(obj)	Returns True if the argument is a dataclass (class or instance).