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
|
package spark
import org.scalatest.FunSuite
import org.scalatest.BeforeAndAfter
import org.scalatest.prop.Checkers
import org.scalacheck.Arbitrary._
import org.scalacheck.Gen
import org.scalacheck.Prop._
import com.google.common.io.Files
import scala.collection.mutable.ArrayBuffer
import SparkContext._
class ShuffleSuite extends FunSuite with BeforeAndAfter {
var sc: SparkContext = _
after {
if (sc != null) {
sc.stop()
sc = null
}
}
test("groupByKey") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1)))
val groups = pairs.groupByKey().collect()
assert(groups.size === 2)
val valuesFor1 = groups.find(_._1 == 1).get._2
assert(valuesFor1.toList.sorted === List(1, 2, 3))
val valuesFor2 = groups.find(_._1 == 2).get._2
assert(valuesFor2.toList.sorted === List(1))
}
test("groupByKey with duplicates") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1)))
val groups = pairs.groupByKey().collect()
assert(groups.size === 2)
val valuesFor1 = groups.find(_._1 == 1).get._2
assert(valuesFor1.toList.sorted === List(1, 1, 2, 3))
val valuesFor2 = groups.find(_._1 == 2).get._2
assert(valuesFor2.toList.sorted === List(1))
}
test("groupByKey with negative key hash codes") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((-1, 1), (-1, 2), (-1, 3), (2, 1)))
val groups = pairs.groupByKey().collect()
assert(groups.size === 2)
val valuesForMinus1 = groups.find(_._1 == -1).get._2
assert(valuesForMinus1.toList.sorted === List(1, 2, 3))
val valuesFor2 = groups.find(_._1 == 2).get._2
assert(valuesFor2.toList.sorted === List(1))
}
test("groupByKey with many output partitions") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (2, 1)))
val groups = pairs.groupByKey(10).collect()
assert(groups.size === 2)
val valuesFor1 = groups.find(_._1 == 1).get._2
assert(valuesFor1.toList.sorted === List(1, 2, 3))
val valuesFor2 = groups.find(_._1 == 2).get._2
assert(valuesFor2.toList.sorted === List(1))
}
test("reduceByKey") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1)))
val sums = pairs.reduceByKey(_+_).collect()
assert(sums.toSet === Set((1, 7), (2, 1)))
}
test("reduceByKey with collectAsMap") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1)))
val sums = pairs.reduceByKey(_+_).collectAsMap()
assert(sums.size === 2)
assert(sums(1) === 7)
assert(sums(2) === 1)
}
test("reduceByKey with many output partitons") {
sc = new SparkContext("local", "test")
val pairs = sc.parallelize(Array((1, 1), (1, 2), (1, 3), (1, 1), (2, 1)))
val sums = pairs.reduceByKey(_+_, 10).collect()
assert(sums.toSet === Set((1, 7), (2, 1)))
}
test("join") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1)))
val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w')))
val joined = rdd1.join(rdd2).collect()
assert(joined.size === 4)
assert(joined.toSet === Set(
(1, (1, 'x')),
(1, (2, 'x')),
(2, (1, 'y')),
(2, (1, 'z'))
))
}
test("join all-to-all") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (1, 3)))
val rdd2 = sc.parallelize(Array((1, 'x'), (1, 'y')))
val joined = rdd1.join(rdd2).collect()
assert(joined.size === 6)
assert(joined.toSet === Set(
(1, (1, 'x')),
(1, (1, 'y')),
(1, (2, 'x')),
(1, (2, 'y')),
(1, (3, 'x')),
(1, (3, 'y'))
))
}
test("leftOuterJoin") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1)))
val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w')))
val joined = rdd1.leftOuterJoin(rdd2).collect()
assert(joined.size === 5)
assert(joined.toSet === Set(
(1, (1, Some('x'))),
(1, (2, Some('x'))),
(2, (1, Some('y'))),
(2, (1, Some('z'))),
(3, (1, None))
))
}
test("rightOuterJoin") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1)))
val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w')))
val joined = rdd1.rightOuterJoin(rdd2).collect()
assert(joined.size === 5)
assert(joined.toSet === Set(
(1, (Some(1), 'x')),
(1, (Some(2), 'x')),
(2, (Some(1), 'y')),
(2, (Some(1), 'z')),
(4, (None, 'w'))
))
}
test("join with no matches") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1)))
val rdd2 = sc.parallelize(Array((4, 'x'), (5, 'y'), (5, 'z'), (6, 'w')))
val joined = rdd1.join(rdd2).collect()
assert(joined.size === 0)
}
test("join with many output partitions") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1)))
val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w')))
val joined = rdd1.join(rdd2, 10).collect()
assert(joined.size === 4)
assert(joined.toSet === Set(
(1, (1, 'x')),
(1, (2, 'x')),
(2, (1, 'y')),
(2, (1, 'z'))
))
}
test("groupWith") {
sc = new SparkContext("local", "test")
val rdd1 = sc.parallelize(Array((1, 1), (1, 2), (2, 1), (3, 1)))
val rdd2 = sc.parallelize(Array((1, 'x'), (2, 'y'), (2, 'z'), (4, 'w')))
val joined = rdd1.groupWith(rdd2).collect()
assert(joined.size === 4)
assert(joined.toSet === Set(
(1, (ArrayBuffer(1, 2), ArrayBuffer('x'))),
(2, (ArrayBuffer(1), ArrayBuffer('y', 'z'))),
(3, (ArrayBuffer(1), ArrayBuffer())),
(4, (ArrayBuffer(), ArrayBuffer('w')))
))
}
test("zero-partition RDD") {
sc = new SparkContext("local", "test")
val emptyDir = Files.createTempDir()
val file = sc.textFile(emptyDir.getAbsolutePath)
assert(file.splits.size == 0)
assert(file.collect().toList === Nil)
// Test that a shuffle on the file works, because this used to be a bug
assert(file.map(line => (line, 1)).reduceByKey(_ + _).collect().toList === Nil)
}
}
|
