Generators

Generator is a function which is responsible to generate a sequence of values.

We can write generator functions just like ordinary functions, but it uses yield keyword to

return values.

Eg 1:

1) def mygen():
2) yield 'A'
3) yield 'B'
4) yield 'C'

5)

6) g=mygen()
7) print(type(g))

8)

Eg 2:

1) def countdown(num):
2) print("Start Countdown")
3) while(num>0):
4) yield num
5) num=num-1

6)

7) values=countdown(5)
8) for x in values:

Generator

Function A Sequence of Values

yield

Eg 3: To generate first n numbers:

1) def firstn(num):
2) n=1
3) while n<=num:
4) yield n
5) n=n+1

6)

We can convert generator into list as follows:

values=firstn(10)

l1=list(values)

print(l1) #[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

Eg 4: To generate Fibonacci Numbers...

The next is the sum of previous 2 numbers

Eg: 0,1,1,2,3,5,8,13,21,...

Advantages of Generator Functions:

1. when compared with class level iterators, generators are very easy to use

• Improves memory utilization and performance.
• Generators are best suitable for reading data from large number of large files
• Generators work great for web scraping and crawling. Generators vs Normal Collections wrt performance:

1) import random
2) import time

3)

4) names = ['Sunny','Bunny','Chinny','Vinny']
5) subjects = ['Python','Java','Blockchain']

6)

7) def people_list(num_people):
8) results = []
9) for i in range(num_people):
10) person = {
11) 'id':i,
12) 'name': random.choice(names),
13) 'subject':random.choice(subjects)
14) }
15) results.append(person)
16) return results

17)

18) def people_generator(num_people):
19) for i in range(num_people):
20) person = {
21) 'id':i,
22) 'name': random.choice(names),
23) 'major':random.choice(subjects)
24) }
25) yield person

26)

27) '''''t1 = time.clock()
28) people = people_list(10000000)
29) t2 = time.clock()'''

30)

31) t1 = time.clock()
32) people = people_generator(10000000)
33) t2 = time.clock()

34)

35) print('Took {}'.format(t2-t1))

Note: In the above program observe the differnce wrt execution time by using list and generators

Generators vs Normal Collections wrt Memory Utilization:

Normal Collection:

l=[x*x for x in range(10000000000000000)]

print(l[0])

We will get MemoryError in this case because all these values are required to store in the memory.

Generators:

g=(x*x for x in range(10000000000000000))

We won't get any MemoryError because the values won't be stored at the beginning