-
Notifications
You must be signed in to change notification settings - Fork 471
/
pyspark-session-2020-10-15.txt
executable file
·221 lines (215 loc) · 7.08 KB
/
pyspark-session-2020-10-15.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
$ cat /Users/mparsian/salaries.txt
alex,sales,20000
bob,business,30000
alex,sales2,25000
bob,business2,35000
alex,sales,26000
bob,business,36000
alex,sales4,21000
bob,business,31000
alex,sales,24000
bob,business,34000
alex,sales6,12000
$ ~/spark-3.0.0/bin/pyspark
Python 3.7.2 (v3.7.2:9a3ffc0492, Dec 24 2018, 02:44:43)
[Clang 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
Using Spark's default log4j profile: org/apache/spark/log4j-defaults.properties
Setting default log level to "WARN".
To adjust logging level use sc.setLogLevel(newLevel). For SparkR, use setLogLevel(newLevel).
Welcome to
____ __
/ __/__ ___ _____/ /__
_\ \/ _ \/ _ `/ __/ '_/
/__ / .__/\_,_/_/ /_/\_\ version 3.0.0
/_/
Using Python version 3.7.2 (v3.7.2:9a3ffc0492, Dec 24 2018 02:44:43)
SparkSession available as 'spark'.
>>>
>>>
>>> # Transformations: map(), flatMap(), mapPartitions(),
... # Reductions: reduce(), reduceByKey(), groupByKey()
...
>>>
>>> nums = [1, 2, 3, 44, 55, 66, 77, 88, 4, 3, 5, 7, 7, 7, 9]
>>> rdd = spark.sparkContext.parallelize(nums)
>>>
>>>
>>> rdd.collect()
[1, 2, 3, 44, 55, 66, 77, 88, 4, 3, 5, 7, 7, 7, 9]
>>> rdd.count()
15
>>> number_of_partitions= rdd.getNumPartitions()
>>> number_of_partitions
8
>>>
>>>
>>>
>>>
>>> def debug(iterator):
... elements = []
... for x in iterator:
... elements.append(x)
... print("elements="+ str(elements))
...
>>>
>>> rdd.foreachPartition(debug)
elements=[44, 55]
elements=[7, 9]
elements=[3, 5]
elements=[7, 7]
elements=[1]
elements=[2, 3]
elements=[88, 4]
elements=[66, 77]
>>> rdd = spark.sparkContext.parallelize(nums)
>>> number_of_partitions= rdd.getNumPartitions()
>>> number_of_partitions
8
>>>
>>> rdd2 = spark.sparkContext.parallelize(nums, 3)
>>> number_of_partitions2 = rdd2.getNumPartitions()
>>> number_of_partitions2
3
>>> rdd2.foreachPartition(debug)
elements=[1, 2, 3, 44, 55]
elements=[66, 77, 88, 4, 3]
elements=[5, 7, 7, 7, 9]
>>>
>>> rdd4 = spark.sparkContext.parallelize(nums, 4)
>>> number_of_partitions4 = rdd4.getNumPartitions()
>>> number_of_partitions4
4
>>> rdd4.foreachPartition(debug)
elements=[1, 2, 3]
elements=[3, 5, 7, 7, 7, 9]
elements=[77, 88, 4]
elements=[44, 55, 66]
>>>
>>>
>>> rdd.foreachPartition(lambda iterator : debug(iterator))
elements=[44, 55]
elements=[7, 9]
elements=[88, 4]
elements=[7, 7]
elements=[2, 3]
elements=[66, 77]
elements=[3, 5]
elements=[1]
>>>
>>>
>>> rdd.collect()
[1, 2, 3, 44, 55, 66, 77, 88, 4, 3, 5, 7, 7, 7, 9]
>>> def myfun(x):
... return (x, x+5)
...
>>>
>>> myfunc(8)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
NameError: name 'myfunc' is not defined
>>> myfun(8)
(8, 13)
>>>
>>>
>>> myfun(90)
(90, 95)
>>> rdd2 = rdd.map(lambda x: myfun(x))
>>> rdd2.collect()
[(1, 6), (2, 7), (3, 8), (44, 49), (55, 60), (66, 71), (77, 82), (88, 93), (4, 9), (3, 8), (5, 10), (7, 12), (7, 12), (7, 12), (9, 14)]
>>>
>>>
>>> rdd3 = rdd.map(myfun)
>>> rdd3.collect()
[(1, 6), (2, 7), (3, 8), (44, 49), (55, 60), (66, 71), (77, 82), (88, 93), (4, 9), (3, 8), (5, 10), (7, 12), (7, 12), (7, 12), (9, 14)]
>>>
>>>
>>>
>>> input_path = "/Users/mparsian/salaries.txt"
>>> input_path
'/Users/mparsian/salaries.txt'
>>>
>>>
>>> # create recs :RDD[String]
...
>>> recs = spark.sparkContext.textFile(input_path)
>>> recs
/Users/mparsian/salaries.txt MapPartitionsRDD[13] at textFile at NativeMethodAccessorImpl.java:0
>>> recs.collect()
['alex,sales,20000', 'bob,business,30000', 'alex,sales2,25000', 'bob,business2,35000', 'alex,sales,26000', 'bob,business,36000', 'alex,sales4,21000', 'bob,business,31000', 'alex,sales,24000', 'bob,business,34000', 'alex,sales6,12000']
>>>
>>>
>>> def create_name_salary_pair(rec):
... tokens = rec.split(",")
... name = tokens[0]
... salary = tokens[2]
... return (name, salary)
...
>>>
>>> create_name_salary_pair('bob,business,34000')
('bob', '34000')
>>> create_name_salary_pair('alex,sales,20000')
('alex', '20000')
>>>
>>>
>>> pairs = recs.map(lambda x: create_name_salary_pair(x))
>>> pairs.collect()
[('alex', '20000'), ('bob', '30000'), ('alex', '25000'), ('bob', '35000'), ('alex', '26000'), ('bob', '36000'), ('alex', '21000'), ('bob', '31000'), ('alex', '24000'), ('bob', '34000'), ('alex', '12000')]
>>> # ('alex', (6, sum, avg), ...
...
>>>
>>> grouped = pairs.groupByKey()
>>>
>>> grouped.mapValues(lambda salaries: list(salaries)).collect()
[('alex', ['20000', '25000', '26000', '21000', '24000', '12000']), ('bob', ['30000', '35000', '36000', '31000', '34000'])]
>>>
>>> def create_name_salary_pair(rec):
... tokens = rec.split(",")
... name = tokens[0]
... salary = tokens[2]
... return (name, int(salary))
...
>>>
>>> pairs = recs.map(lambda x: create_name_salary_pair(x))
>>> pairs.collect()
[('alex', 20000), ('bob', 30000), ('alex', 25000), ('bob', 35000), ('alex', 26000), ('bob', 36000), ('alex', 21000), ('bob', 31000), ('alex', 24000), ('bob', 34000), ('alex', 12000)]
>>> grouped = pairs.groupByKey()
>>> grouped.collect()
[('alex', <pyspark.resultiterable.ResultIterable object at 0x7ffc961fb3c8>), ('bob', <pyspark.resultiterable.ResultIterable object at 0x7ffc961fb400>)]
>>> grouped.mapValues(lambda salaries: list(salaries)).collect()
[('alex', [20000, 25000, 26000, 21000, 24000, 12000]), ('bob', [30000, 35000, 36000, 31000, 34000])]
>>>
>>> final_rdd = grouped.mapValues(lambda salaries: (len(salaries), sum(salaries), sum(salaries)/len(salaries)))
>>> final_rdd.collect()
[('alex', (6, 128000, 21333.333333333332)), ('bob', (5, 166000, 33200.0))]
>>>
>>>
>>> final_rdd2 = grouped.map(lambda x: (x[0], (len(x[1]), sum(x[1]), sum(x[1])/len(x[1]))))
>>> final_rdd2.collect()
[('alex', (6, 128000, 21333.333333333332)), ('bob', (5, 166000, 33200.0))]
>>>
>>> recs.collect()
['alex,sales,20000', 'bob,business,30000', 'alex,sales2,25000', 'bob,business2,35000', 'alex,sales,26000', 'bob,business,36000', 'alex,sales4,21000', 'bob,business,31000', 'alex,sales,24000', 'bob,business,34000', 'alex,sales6,12000']
>>> pairs = recs.map(lambda x: create_name_salary_pair(x))
>>> pairs.collect()
[('alex', 20000), ('bob', 30000), ('alex', 25000), ('bob', 35000), ('alex', 26000), ('bob', 36000), ('alex', 21000), ('bob', 31000), ('alex', 24000), ('bob', 34000), ('alex', 12000)]
>>>
>>> tuples2 = pairs.mapValues(lambda s : (s, 1))
>>> tuples2.collect()
[('alex', (20000, 1)), ('bob', (30000, 1)), ('alex', (25000, 1)), ('bob', (35000, 1)), ('alex', (26000, 1)), ('bob', (36000, 1)), ('alex', (21000, 1)), ('bob', (31000, 1)), ('alex', (24000, 1)), ('bob', (34000, 1)), ('alex', (12000, 1))]
>>>
>>>
>>> reduced = tuples2.reduceByKey(lambda a, b: (a[0]+b[0], a[1]+b[1]))
>>> reduced.collect()
[('alex', (128000, 6)), ('bob', (166000, 5))]
>>>
>>> final_answer_by_reducebykey = reduced.mapValues(lambda v : (v[1], v[0], v[0]/v[1]))
>>> final_answer_by_reducebykey.collect()
[('alex', (6, 128000, 21333.333333333332)), ('bob', (5, 166000, 33200.0))]
>>>
>>>
>>> # find (avg, median) per key
... # 1. by using groupByKey()
... # 2. by using reduceByKey()
...
>>>