""" Day 43 - 03-dataclass-vs-namedtuple.py dataclass 与 namedtuple / 普通类的对比 本文件通过实际对比展示三种定义数据容器的方式的差异， 帮助你在不同场景下做出合适的选择。 """ from dataclasses import dataclass, field, fields from typing import NamedTuple, List, Optional from collections import namedtuple import time import sys # ============================================================ # 1. 相同功能，三种实现 # ============================================================ # --- 普通类 --- # 需要手写所有模板方法 class NormalPerson: """普通类实现——需要写大量模板代码""" def __init__(self, name: str, age: int, email: str = ""): self.name = name self.age = age self.email = email def __repr__(self) -> str: return f"NormalPerson(name={self.name!r}, age={self.age!r}, email={self.email!r})" def __eq__(self, other: object) -> bool: if not isinstance(other, NormalPerson): return NotImplemented return (self.name, self.age, self.email) == (other.name, other.age, other.email) # --- namedtuple --- # 最简洁，但功能有限 # 经典方式 PersonTuple = namedtuple("PersonTuple", ["name", "age", "email"]) # 默认值——经典版的 default 参数 (Python 3.7+) PersonTupleDefault = namedtuple("PersonTupleDefault", ["name", "age", "email"], defaults=[""]) # 现代方式：用 typing.NamedTuple（支持类型注解和方法） class PersonNamedTuple(NamedTuple): name: str age: int email: str = "" def greet(self) -> str: return f"你好，我是 {self.name}" # --- dataclass --- # 简洁且功能丰富 @dataclass class PersonDataClass: """数据类实现——几行代码搞定""" name: str age: int email: str = "" def greet(self) -> str: return f"你好，我是 {self.name}" def demo_creation_comparison() -> None: """对比三种方式的代码量和易用性""" print("=" * 60) print("1. 代码量对比") print("=" * 60) # 普通类 np = NormalPerson("Alice", 30, "alice@example.com") print(f"普通类: {np} | {np.greet() if hasattr(np, 'greet') else '无方法'}") # namedtuple nt = PersonNamedTuple("Bob", 25, "bob@example.com") print(f"namedtuple: {nt} | {nt.greet()}") # dataclass dc = PersonDataClass("Charlie", 35, "charlie@example.com") print(f"dataclass: {dc} | {dc.greet()}") print() # ============================================================ # 2. 可变性对比 # ============================================================ def demo_mutability() -> None: """对比三种方式的不可变性""" print("=" * 60) print("2. 可变性对比") print("=" * 60) # ★ dataclass 默认可变 dc = PersonDataClass("Alice", 30, "alice@example.com") dc.age = 31 # ✅ 可以修改 print(f"dataclass 可变: {dc}") # ★ dataclass frozen=True 不可变 @dataclass(frozen=True) class FrozenPerson: name: str age: int fp = FrozenPerson("Bob", 25) print(f"dataclass frozen: {fp}") try: fp.age = 26 except Exception as e: print(f" 尝试修改 ➜ {type(e).__name__}: {e}") # ★ namedtuple 不可变 nt = PersonNamedTuple("Charlie", 35, "charlie@example.com") print(f"namedtuple: {nt}") try: nt.age = 36 except Exception as e: print(f" 尝试修改 ➜ {type(e).__name__}: {e}") # ★ 普通类 可变 nc = NormalPerson("Diana", 28) nc.age = 29 # ✅ print(f"普通类可变: {nc}") print() # ============================================================ # 3. 访问方式对比 # ============================================================ def demo_access() -> None: """对比三种方式的字段访问方式""" print("=" * 60) print("3. 字段访问方式") print("=" * 60) # 属性访问：三种都支持 dc = PersonDataClass("Alice", 30) nt = PersonNamedTuple("Bob", 25) nc = NormalPerson("Charlie", 35) # 属性访问（点语法）— 都支持 print(f"属性访问 — dataclass: {dc.name}") print(f"属性访问 — namedtuple: {nt.name}") print(f"属性访问 — 普通类: {nc.name}") # 索引访问 — 只有 namedtuple 支持 print(f"\n索引访问 — namedtuple[0]: {nt[0]}") # Bob print(f"索引访问 — namedtuple[1]: {nt[1]}") # 25 # 拆包 — 只有 namedtuple 支持 name, age, email = nt print(f"\n拆包 — namedtuple: name={name}, age={age}") try: print(dc[0]) # ❌ except TypeError as e: print(f"dataclass 不支持索引: {e}") # 转字典 import dataclasses print(f"\ndataclasses.asdict: {dataclasses.asdict(dc)}") print(f"namedtuple._asdict: {nt._asdict()}") print(f"vars(普通类): {vars(nc)}") print() # ============================================================ # 4. 性能对比 # ============================================================ def demo_performance() -> None: """三种方式的创建性能对比""" print("=" * 60) print("4. 创建性能对比（创建 100 万个对象）") print("=" * 60) # 预热 N = 1_000_000 # dataclass 创建 start = time.perf_counter() dcs = [PersonDataClass(f"User-{i}", i % 100) for i in range(N)] dc_time = time.perf_counter() - start # namedtuple 创建 start = time.perf_counter() nts = [PersonNamedTuple(f"User-{i}", i % 100) for i in range(N)] nt_time = time.perf_counter() - start # 普通类创建 start = time.perf_counter() ncs = [NormalPerson(f"User-{i}", i % 100) for i in range(N)] nc_time = time.perf_counter() - start print(f" dataclass: {dc_time:.3f} 秒 ({dc_time/nc_time:.2f}x 普通类)") print(f" namedtuple: {nt_time:.3f} 秒 ({nt_time/nc_time:.2f}x 普通类)") print(f" 普通类: {nc_time:.3f} 秒 (基准)") # 读取性能 start = time.perf_counter() for p in dcs: _ = p.name, p.age dc_read = time.perf_counter() - start start = time.perf_counter() for p in nts: _ = p.name, p.age nt_read = time.perf_counter() - start print(f"\n属性读取性能（100 万次）:") print(f" dataclass: {dc_read:.3f} 秒") print(f" namedtuple: {nt_read:.3f} 秒") # 内存对比 print(f"\n内存占用（取 1 个实例的粗略大小）:") print(f" dataclass: {sys.getsizeof(dcs[0])} 字节") print(f" namedtuple: {sys.getsizeof(nts[0])} 字节") print(f" 普通类: {sys.getsizeof(ncs[0])} 字节") print() # ============================================================ # 5. 功能丰富度对比 # ============================================================ @dataclass class FeatureRich: """dataclass 高级功能演示""" name: str items: List[str] = field(default_factory=list) # ^^^^^^^^^^^^^^^^^^^^^^^ 可变类型默认值 _secret: str = field(default="", repr=False) # ^^^^^^^^^^^^^^^^^^^^^^ 隐藏字段 id: str = field(compare=False, default="") # ^^^^^^^^^^^^^ 排除比较字段 def __post_init__(self) -> None: """初始化后处理""" self.validate() def validate(self) -> None: if not self.name: raise ValueError("名称不能为空") def add_item(self, item: str) -> None: self.items.append(item) # namedtuple 也能添加方法，但很有限 class AdvancedNamedTuple(NamedTuple): """namedtuple 扩展——通过继承添加方法""" name: str items: tuple = () # namedtuple 不可变，只能用元组 def add_item(self, item: str) -> "AdvancedNamedTuple": # 不可变对象只能返回新实例 return self._replace(items=self.items + (item,)) def item_count(self) -> int: return len(self.items) def demo_feature_comparison() -> None: """对比功能丰富度""" print("=" * 60) print("5. 功能丰富度对比") print("=" * 60) # dataclass：可变，支持 field()，__post_init__ dc = FeatureRich("数据包", _secret="hidden") dc.add_item("日志条目1") dc.add_item("日志条目2") print(f"dataclass: {dc}") # repr 隐藏了 _secret print(f" items: {dc.items}") # namedtuple：不可变，修改需要创建新对象 nt = AdvancedNamedTuple("数据包") nt2 = nt.add_item("日志条目1") nt3 = nt2.add_item("日志条目2") print(f"\nnamedtuple: {nt}") # 原对象不变 print(f" add_item 后: {nt3}") print(f" item_count: {nt3.item_count()}") # dataclass 支持默认工厂、metadata 等： print(f"\ndataclass fields 数量: {len(fields(FeatureRich))}") # namedtuple 的不变性使它可哈希 print(f"\nnamedtuple 可哈希: {hasattr(nt, '__hash__') and nt.__hash__ is not None}") print(f"dataclass 可哈希: {hasattr(dc, '__hash__') and dc.__hash__ is not None}") # False（默认） print() # ============================================================ # 6. 继承对比 # ============================================================ @dataclass class Shape: """基类""" name: str def area(self) -> float: return 0.0 @dataclass class Circle(Shape): """dataclass 继承——简洁""" radius: float def area(self) -> float: return 3.14159 * self.radius ** 2 @dataclass class ColoredCircle(Circle): """多层继承""" color: str = "red" class CircleNT(NamedTuple): """圆——NamedTuple 可以直接定义方法""" name: str radius: float def area(self) -> float: return 3.14159 * self.radius ** 2 def demo_inheritance() -> None: """对比继承能力""" print("=" * 60) print("6. 继承对比") print("=" * 60) # dataclass 继承 c = ColoredCircle("小红圆", 5.0, "red") print(f"dataclass 继承: {c}") print(f" area: {c.area():.2f}") print(f" 多层继承: ColoredCircle → Circle → Shape") # NamedTuple 也可以定义方法和计算属性 cn = CircleNT("大蓝圆", 10.0) print(f"\nnamedtuple 方法: {cn}") print(f" area: {cn.area():.2f}") # ⚠️ NamedTuple 子类添加字段需要 Python 3.6+ 的 typing.NamedTuple # 但不支持多层继承添加字段，这是 dataclass 的优势 print() # ============================================================ # 入口 # ============================================================ if __name__ == "__main__": demo_creation_comparison() demo_mutability() demo_access() demo_performance() demo_feature_comparison() demo_inheritance() print("=" * 60) print("🎉 三种方式对比演示完成！") print("=" * 60) print() print("📌 总结建议：") print(" ✅ 日常首选 dataclass — 功能/简洁的最佳平衡") print(" ✅ 不可变 + 性能敏感 → NamedTuple") print(" ✅ 简单元组兼容 + 拆包 → namedtuple") print(" ✅ 复杂业务逻辑 → 手写 class")