[python]yield用法浅析

创建生成器(generator)有两种方法：

• 将list comprehension中的方括号[]变成圆括号()

>>> a=(i for i in range(10))
>>> a
<generator object <genexpr> at 0x02C0A9B8>

• 在函数中使用yield，该函数自动变成一个生成器

>>> def gen():
	print 'generator'
	yield 0
	print 'step1'
	yield 1
	print 'step2'
	yield 2
	print 'Bye!'

	
>>> g=gen()
>>> g
<generator object gen at 0x02C0A788>

生成器的工作原理是，在执行生成器的时候（其实就是执行函数内的代码），遇到yield就立马停止，并返回yield后面的数据，直到yield下一行没有代码，此时抛出“StopIteration”异常，生成器停止工作。以上面定义的gen()为例进行说明。

>>> g=gen()   # 得到生成器g
>>> g
<generator object gen at 0x02C0A788>
>>> g.next()  # 执行生成器直到遇到yield
generator
0
>>> g.next()  # 从上次执行yield命令的下一行继续执行，直到遇到下一个yield
step1
1
>>> g.next()
step2
2
>>> g.next()  # yield后面没有内容，抛出'StopIteration'异常
Bye!

Traceback (most recent call last):
  File "<pyshell#145>", line 1, in <module>
    g.next()
StopIteration

如果将生成器放到for...in...中——for value in generator，则无需人工调用next()方法，当generator(函数执行过程中)遇到yield的时候，会将yield后面的数据返回给for中的value，然后程序会自动执行next()。如果遇到下一个yield，则重复前面的过程；如果程序执行完毕，则退出for循环。

>>> g=gen()
>>> g
<generator object gen at 0x02C0A788>
>>> for i in g:   # 每次将yield后面的数据返回给i
	print '<value^2={}>'.format(i*i)

	
generator
<value^2=0>
step1
<value^2=1>
step2
<value^2=4>
Bye!

举例，下面的例子展示了输出长度为9的链表，链表中的内容是[0],[0,1],[0,1,2],...[0,1,2,...,8]。使用生成器可以不用将所有待打印的数据保存下来，每次执行过程中才生成相应的序列，然后打出输出。节省了内存空间。

>>> def gen(n):
	a=range(n)
	for i in a:
		yield i

		
>>> l=[]
>>> for i in range(1,10):
	l.append(gen(i))

	
>>> l
[<generator object gen at 0x02C4D2D8>, <generator object gen at 0x02C4E440>, <generator object gen at 0x02C4E148>, <generator object gen at 0x02C4E7D8>, <generator object gen at 0x02C4E670>, <generator object gen at 0x02C4EA80>, <generator object gen at 0x02C4E9E0>, <generator object gen at 0x02C4EC60>, <generator object gen at 0x02C4ED00>]
>>> for i in l:
	for j in i:
		print '[{}]value={}'.format(i,j)

		
[<generator object gen at 0x02C4D2D8>]value=0
[<generator object gen at 0x02C4E440>]value=0
[<generator object gen at 0x02C4E440>]value=1
[<generator object gen at 0x02C4E148>]value=0
[<generator object gen at 0x02C4E148>]value=1
[<generator object gen at 0x02C4E148>]value=2
[<generator object gen at 0x02C4E7D8>]value=0
[<generator object gen at 0x02C4E7D8>]value=1
[<generator object gen at 0x02C4E7D8>]value=2
[<generator object gen at 0x02C4E7D8>]value=3
[<generator object gen at 0x02C4E670>]value=0
[<generator object gen at 0x02C4E670>]value=1
[<generator object gen at 0x02C4E670>]value=2
[<generator object gen at 0x02C4E670>]value=3
[<generator object gen at 0x02C4E670>]value=4
[<generator object gen at 0x02C4EA80>]value=0
[<generator object gen at 0x02C4EA80>]value=1
[<generator object gen at 0x02C4EA80>]value=2
[<generator object gen at 0x02C4EA80>]value=3
[<generator object gen at 0x02C4EA80>]value=4
[<generator object gen at 0x02C4EA80>]value=5
[<generator object gen at 0x02C4E9E0>]value=0
[<generator object gen at 0x02C4E9E0>]value=1
[<generator object gen at 0x02C4E9E0>]value=2
[<generator object gen at 0x02C4E9E0>]value=3
[<generator object gen at 0x02C4E9E0>]value=4
[<generator object gen at 0x02C4E9E0>]value=5
[<generator object gen at 0x02C4E9E0>]value=6
[<generator object gen at 0x02C4EC60>]value=0
[<generator object gen at 0x02C4EC60>]value=1
[<generator object gen at 0x02C4EC60>]value=2
[<generator object gen at 0x02C4EC60>]value=3
[<generator object gen at 0x02C4EC60>]value=4
[<generator object gen at 0x02C4EC60>]value=5
[<generator object gen at 0x02C4EC60>]value=6
[<generator object gen at 0x02C4EC60>]value=7
[<generator object gen at 0x02C4ED00>]value=0
[<generator object gen at 0x02C4ED00>]value=1
[<generator object gen at 0x02C4ED00>]value=2
[<generator object gen at 0x02C4ED00>]value=3
[<generator object gen at 0x02C4ED00>]value=4
[<generator object gen at 0x02C4ED00>]value=5
[<generator object gen at 0x02C4ED00>]value=6
[<generator object gen at 0x02C4ED00>]value=7
[<generator object gen at 0x02C4ED00>]value=8

总结如下：

• 得到生成器的时候，并不会执行函数，只有当第一次调用next()方法的时候才会从函数的第一行开始执行
• 在函数内执行代码的过程中，一旦遇到yield关键字就会立马停止运行，并返回yield后面的内容。下次再调用next()方法会激活生成器，并从yield代码的下面一行开始继续执行。如果后面还会遇到yield，重复前面描述的过程。直到函数中的代码被执行完毕，此时会抛出‘StopIteration’异常