Python3 新特性¶
print 是一个函数¶
Python 3.0 新特性
PEP 3105 - 将 print 变成函数
Python 2
>>> print "print is a statement"
print is a statement
>>> for x in range(3):
... print x,
...
0 1 2
Python 3
>>> print("print is a function")
print is a function
>>> print()
>>> for x in range(3):
... print(x, end=' ')
... else:
... print()
...
0 1 2
字符串为 Unicode¶
Python 3.0 新特性
Python 2
>>> s = 'Café' # byte string
>>> s
'Caf\xc3\xa9'
>>> type(s)
<type 'str'>
>>> u = u'Café' # unicode string
>>> u
u'Caf\xe9'
>>> type(u)
<type 'unicode'>
>>> len([_c for _c in 'Café'])
5
Python 3
>>> s = 'Café'
>>> s
'Café'
>>> type(s)
<class 'str'>
>>> s.encode('utf-8')
b'Caf\xc3\xa9'
>>> s.encode('utf-8').decode('utf-8')
'Café'
>>> len([_c for _c in 'Café'])
4
除法运算符¶
Python 3.0 新特性
PEP 238 - 更改除法运算符
Python2
>>> 1 / 2
0
>>> 1 // 2
0
>>> 1. / 2
0.5
# back port "true division" to python2
>>> from __future__ import division
>>> 1 / 2
0.5
>>> 1 // 2
0
Python3
>>> 1 / 2
0.5
>>> 1 // 2
0
新的字典实现¶
Python 3.6 新特性
Python 3.5 之前
>>> import sys
>>> sys.getsizeof({str(i):i for i in range(1000)})
49248
>>> d = {'timmy': 'red', 'barry': 'green', 'guido': 'blue'}
>>> d # without order-preserving
{'barry': 'green', 'timmy': 'red', 'guido': 'blue'}
Python 3.6
内存使用量小于 Python 3.5
保留插入顺序
>>> import sys
>>> sys.getsizeof({str(i):i for i in range(1000)})
36968
>>> d = {'timmy': 'red', 'barry': 'green', 'guido': 'blue'}
>>> d # preserve insertion ordered
{'timmy': 'red', 'barry': 'green', 'guido': 'blue'}
仅限关键字参数¶
Python 3.0 新特性
PEP 3102 - 仅限关键字参数
>>> def f(a, b, *, kw):
... print(a, b, kw)
...
>>> f(1, 2, 3)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: f() takes 2 positional arguments but 3 were given
>>> f(1, 2)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: f() missing 1 required keyword-only argument: 'kw'
>>> f(1, 2, kw=3)
1 2 3
新的 Super¶
Python 3.0 新特性
PEP 3135 - 新的 Super
Python 2
>>> class ParentCls(object):
... def foo(self):
... print "call parent"
...
>>> class ChildCls(ParentCls):
... def foo(self):
... super(ChildCls, self).foo()
... print "call child"
...
>>> p = ParentCls()
>>> c = ChildCls()
>>> p.foo()
call parent
>>> c.foo()
call parent
call child
Python 3
>>> class ParentCls(object):
... def foo(self):
... print("call parent")
...
>>> class ChildCls(ParentCls):
... def foo(self):
... super().foo()
... print("call child")
...
>>> p = ParentCls()
>>> c = ChildCls()
>>> p.foo()
call parent
>>> c.foo()
call parent
call child
移除 <>¶
Python 3.0 新特性
Python 2
>>> a = "Python2"
>>> a <> "Python3"
True
# equal to !=
>>> a != "Python3"
True
Python 3
>>> a = "Python3"
>>> a != "Python2"
True
BDFL 退休¶
Python 3.1 新特性
PEP 401 - BDFL 退休
>>> from __future__ import barry_as_FLUFL
>>> 1 != 2
File "<stdin>", line 1
1 != 2
^
SyntaxError: with Barry as BDFL, use '<>' instead of '!='
>>> 1 <> 2
True
不允许在函数内部使用 from module import *¶
Python 3.0 新特性
>>> def f():
... from os import *
...
File "<stdin>", line 1
SyntaxError: import * only allowed at module level
添加 nonlocal 关键字¶
Python 3.0 新特性
PEP 3104 - 访问外部作用域中的名称
注意
nonlocal 允许直接赋值给外部（但非全局）作用域中的变量
>>> def outf():
... o = "out"
... def inf():
... nonlocal o
... o = "change out"
... inf()
... print(o)
...
>>> outf()
change out
扩展的可迭代解包¶
Python 3.0 新特性
PEP 3132 - 扩展的可迭代解包
>>> a, *b, c = range(5)
>>> a, b, c
(0, [1, 2, 3], 4)
>>> for a, *b in [(1, 2, 3), (4, 5, 6, 7)]:
... print(a, b)
...
1 [2, 3]
4 [5, 6, 7]
通用解包¶
Python 3.5 新特性
PEP 448 - 额外的解包泛化
Python 2
>>> def func(*a, **k):
... print(a)
... print(k)
...
>>> func(*[1,2,3,4,5], **{"foo": "bar"})
(1, 2, 3, 4, 5)
{'foo': 'bar'}
Python 3
>>> print(*[1, 2, 3], 4, *[5, 6])
1 2 3 4 5 6
>>> [*range(4), 4]
[0, 1, 2, 3, 4]
>>> {"foo": "Foo", "bar": "Bar", **{"baz": "baz"}}
{'foo': 'Foo', 'bar': 'Bar', 'baz': 'baz'}
>>> def func(*a, **k):
... print(a)
... print(k)
...
>>> func(*[1], *[4,5], **{"foo": "FOO"}, **{"bar": "BAR"})
(1, 4, 5)
{'foo': 'FOO', 'bar': 'BAR'}
函数注解¶
Python 3.0 新特性
>>> import types
>>> generator = types.GeneratorType
>>> def fib(n: int) -> generator:
... a, b = 0, 1
... for _ in range(n):
... yield a
... b, a = a + b, b
...
>>> [f for f in fib(10)]
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]
变量注解¶
Python 3.6 新特性
PEP 526 - 变量注解语法
>>> from typing import List
>>> x: List[int] = [1, 2, 3]
>>> x
[1, 2, 3]
>>> from typing import List, Dict
>>> class Cls(object):
... x: List[int] = [1, 2, 3]
... y: Dict[str, str] = {"foo": "bar"}
...
>>> o = Cls()
>>> o.x
[1, 2, 3]
>>> o.y
{'foo': 'bar'}
核心支持 typing 模块和泛型¶
Python 3.7 新特性
PEP 560 - 核心支持 typing 模块和泛型
Python 3.7 之前
>>> from typing import Generic, TypeVar
>>> from typing import Iterable
>>> T = TypeVar('T')
>>> class C(Generic[T]): ...
...
>>> def func(l: Iterable[C[int]]) -> None:
... for i in l:
... print(i)
...
>>> func([1,2,3])
1
2
3
Python 3.7 或更高版本
>>> from typing import Iterable
>>> class C:
... def __class_getitem__(cls, item):
... return f"{cls.__name__}[{item.__name__}]"
...
>>> def func(l: Iterable[C[int]]) -> None:
... for i in l:
... print(i)
...
>>> func([1,2,3])
1
2
3
格式化字节字符串¶
Python 3.5 新特性
PEP 461 - 为 bytes 和 bytearray 添加
%格式化
>>> b'abc %b %b' % (b'foo', b'bar')
b'abc foo bar'
>>> b'%d %f' % (1, 3.14)
b'1 3.140000'
>>> class Cls(object):
... def __repr__(self):
... return "repr"
... def __str__(self):
... return "str"
...
'repr'
>>> b'%a' % Cls()
b'repr'
f 字符串¶
Python 3.6 新特性
PEP 498 - 字面量字符串插值
>>> py = "Python3"
>>> f'Awesome {py}'
'Awesome Python3'
>>> x = [1, 2, 3, 4, 5]
>>> f'{x}'
'[1, 2, 3, 4, 5]'
>>> def foo(x:int) -> int:
... return x + 1
...
>>> f'{foo(0)}'
'1'
>>> f'{123.567:1.3}'
'1.24e+02'
抑制异常¶
Python 3.3 新特性
PEP 409 - 抑制异常上下文
不使用 raise Exception from None
>>> def func():
... try:
... 1 / 0
... except ZeroDivisionError:
... raise ArithmeticError
...
>>> func()
Traceback (most recent call last):
File "<stdin>", line 3, in func
ZeroDivisionError: division by zero
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "<stdin>", line 5, in func
ArithmeticError
使用 raise Exception from None
>>> def func():
... try:
... 1 / 0
... except ZeroDivisionError:
... raise ArithmeticError from None
...
>>> func()
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "<stdin>", line 5, in func
ArithmeticError
# debug
>>> try:
... func()
... except ArithmeticError as e:
... print(e.__context__)
...
division by zero
生成器委托¶
Python 3.3 新特性
PEP 380 - 委托给子生成器的语法
>>> def fib(n: int):
... a, b = 0, 1
... for _ in range(n):
... yield a
... b, a = a + b, b
...
>>> def delegate(n: int):
... yield from fib(n)
...
>>> list(delegate(10))
[0, 1, 1, 2, 3, 5, 8, 13, 21, 34]
async 和 await 语法¶
Python 3.5 新特性
PEP 492 - 使用 async 和 await 语法的协程
Python 3.5 之前
>>> import asyncio
>>> @asyncio.coroutine
... def fib(n: int):
... a, b = 0, 1
... for _ in range(n):
... b, a = a + b, b
... return a
...
>>> @asyncio.coroutine
... def coro(n: int):
... for x in range(n):
... yield from asyncio.sleep(1)
... f = yield from fib(x)
... print(f)
...
>>> loop = asyncio.get_event_loop()
>>> loop.run_until_complete(coro(3))
0
1
1
Python 3.5 或更高版本
>>> import asyncio
>>> async def fib(n: int):
... a, b = 0, 1
... for _ in range(n):
... b, a = a + b, b
... return a
...
>>> async def coro(n: int):
... for x in range(n):
... await asyncio.sleep(1)
... f = await fib(x)
... print(f)
...
>>> loop = asyncio.get_event_loop()
>>> loop.run_until_complete(coro(3))
0
1
1
异步生成器¶
Python 3.6 新特性
PEP 525 - 异步生成器
>>> import asyncio
>>> async def fib(n: int):
... a, b = 0, 1
... for _ in range(n):
... await asyncio.sleep(1)
... yield a
... b, a = a + b , b
...
>>> async def coro(n: int):
... ag = fib(n)
... f = await ag.asend(None)
... print(f)
... f = await ag.asend(None)
... print(f)
...
>>> loop = asyncio.get_event_loop()
>>> loop.run_until_complete(coro(5))
0
1
异步推导式¶
Python 3.6 新特性
PEP 530 - 异步推导式
>>> import asyncio
>>> async def fib(n: int):
... a, b = 0, 1
... for _ in range(n):
... await asyncio.sleep(1)
... yield a
... b, a = a + b , b
...
# async for ... else
>>> async def coro(n: int):
... async for f in fib(n):
... print(f, end=" ")
... else:
... print()
...
>>> loop = asyncio.get_event_loop()
>>> loop.run_until_complete(coro(5))
0 1 1 2 3
# async for in list
>>> async def coro(n: int):
... return [f async for f in fib(n)]
...
>>> loop.run_until_complete(coro(5))
[0, 1, 1, 2, 3]
# await in list
>>> async def slowfmt(n: int) -> str:
... await asyncio.sleep(0.5)
... return f'{n}'
...
>>> async def coro(n: int):
... return [await slowfmt(f) async for f in fib(n)]
...
>>> loop.run_until_complete(coro(5))
['0', '1', '1', '2', '3']
矩阵乘法¶
Python 3.5 新特性
PEP 465 - 用于矩阵乘法的专用中缀运算符
>>> # "@" represent matrix multiplication
>>> class Arr:
... def __init__(self, *arg):
... self._arr = arg
... def __matmul__(self, other):
... if not isinstance(other, Arr):
... raise TypeError
... if len(self) != len(other):
... raise ValueError
... return sum([x*y for x, y in zip(self._arr, other._arr)])
... def __imatmul__(self, other):
... if not isinstance(other, Arr):
... raise TypeError
... if len(self) != len(other):
... raise ValueError
... res = sum([x*y for x, y in zip(self._arr, other._arr)])
... self._arr = [res]
... return self
... def __len__(self):
... return len(self._arr)
... def __str__(self):
... return self.__repr__()
... def __repr__(self):
... return "Arr({})".format(repr(self._arr))
...
>>> a = Arr(9, 5, 2, 7)
>>> b = Arr(5, 5, 6, 6)
>>> a @ b # __matmul__
124
>>> a @= b # __imatmul__
>>> a
Arr([124])
数据类¶
Python 3.7 新特性
PEP 557 - 数据类
可变数据类
>>> from dataclasses import dataclass
>>> @dataclass
... class DCls(object):
... x: str
... y: str
...
>>> d = DCls("foo", "bar")
>>> d
DCls(x='foo', y='bar')
>>> d = DCls(x="foo", y="baz")
>>> d
DCls(x='foo', y='baz')
>>> d.z = "bar"
不可变数据类
>>> from dataclasses import dataclass
>>> from dataclasses import FrozenInstanceError
>>> @dataclass(frozen=True)
... class DCls(object):
... x: str
... y: str
...
>>> try:
... d.x = "baz"
... except FrozenInstanceError as e:
... print(e)
...
cannot assign to field 'x'
>>> try:
... d.z = "baz"
... except FrozenInstanceError as e:
... print(e)
...
cannot assign to field 'z'
内置 breakpoint()¶
Python 3.7 新特性
PEP 553 - 内置 breakpoint()
>>> for x in range(3):
... print(x)
... breakpoint()
...
0
> <stdin>(1)<module>()->None
(Pdb) c
1
> <stdin>(1)<module>()->None
(Pdb) c
2
> <stdin>(1)<module>()->None
(Pdb) c
海象运算符¶
Python 3.8 新特性
PEP 572 - 赋值表达式
海象运算符的目的是在表达式中赋值变量。完成 PEP 572 后，通常被称为 BDFL 的 Guido van Rossum 决定辞去 Python 核心开发者一职。
>>> f = (0, 1)
>>> [(f := (f[1], sum(f)))[0] for i in range(10)]
[1, 1, 2, 3, 5, 8, 13, 21, 34, 55]
仅限位置参数¶
Python 3.8 新特性
PEP 570 - Python 仅限位置参数
>>> def f(a, b, /, c, d):
... print(a, b, c, d)
...
>>> f(1, 2, 3, 4)
1 2 3 4
>>> f(1, 2, c=3, d=4)
1 2 3 4
>>> f(1, b=2, c=3, d=4)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: f() got some positional-only arguments passed as keyword arguments: 'b'
字典合并¶
Python 3.9 新特性
PEP 584 - 为 dict 添加联合运算符
>>> a = {"foo": "Foo"}
>>> b = {"bar": "Bar"}
# old way
>>> {**a, **b}
{'foo': 'Foo', 'bar': 'Bar'}
>>> a.update(b)
>>> a
{'foo': 'Foo', 'bar': 'Bar'}
# new way
>>> a | b
{'foo': 'Foo', 'bar': 'Bar'}
>>> a |= b
>>> a
{'foo': 'Foo', 'bar': 'Bar'}
