-
Notifications
You must be signed in to change notification settings - Fork 28k
/
QueryTest.scala
458 lines (410 loc) · 16 KB
/
QueryTest.scala
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
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.spark.sql
import java.util.TimeZone
import java.util.regex.Pattern
import scala.jdk.CollectionConverters._
import org.scalatest.Assertions
import org.apache.spark.sql.catalyst.ExtendedAnalysisException
import org.apache.spark.sql.catalyst.plans._
import org.apache.spark.sql.catalyst.util._
import org.apache.spark.sql.execution.SQLExecution
import org.apache.spark.sql.execution.columnar.InMemoryRelation
import org.apache.spark.storage.StorageLevel
import org.apache.spark.util.ArrayImplicits._
abstract class QueryTest extends PlanTest {
protected def spark: SparkSession
/**
* Runs the plan and makes sure the answer contains all of the keywords.
*/
def checkKeywordsExist(df: DataFrame, keywords: String*): Unit = {
val outputs = df.collect().map(_.mkString).mkString
for (key <- keywords) {
assert(outputs.contains(key), s"Failed for $df ($key doesn't exist in result)")
}
}
/**
* Runs the plan and makes sure the answer does NOT contain any of the keywords.
*/
def checkKeywordsNotExist(df: DataFrame, keywords: String*): Unit = {
val outputs = df.collect().map(_.mkString).mkString
for (key <- keywords) {
assert(!outputs.contains(key), s"Failed for $df ($key existed in the result)")
}
}
/**
* Evaluates a dataset to make sure that the result of calling collect matches the given
* expected answer.
*/
protected def checkDataset[T](
ds: => Dataset[T],
expectedAnswer: T*): Unit = {
val result = getResult(ds)
if (!QueryTest.compare(result.toSeq, expectedAnswer)) {
fail(
s"""
|Decoded objects do not match expected objects:
|expected: $expectedAnswer
|actual: ${result.toSeq}
|${ds.exprEnc.deserializer.treeString}
""".stripMargin)
}
}
/**
* Evaluates a dataset to make sure that the result of calling collect matches the given
* expected answer, after sort.
*/
protected def checkDatasetUnorderly[T : Ordering](
ds: => Dataset[T],
expectedAnswer: T*): Unit = {
val result = getResult(ds)
if (!QueryTest.compare(result.toSeq.sorted, expectedAnswer.sorted)) {
fail(
s"""
|Decoded objects do not match expected objects:
|expected: $expectedAnswer
|actual: ${result.toSeq}
|${ds.exprEnc.deserializer.treeString}
""".stripMargin)
}
}
private def getResult[T](ds: => Dataset[T]): Array[T] = {
val analyzedDS = try ds catch {
case ae: ExtendedAnalysisException =>
if (ae.plan.isDefined) {
fail(
s"""
|Failed to analyze query: $ae
|${ae.plan.get}
|
|${stackTraceToString(ae)}
""".stripMargin)
} else {
throw ae
}
}
assertEmptyMissingInput(analyzedDS)
try ds.collect() catch {
case e: Exception =>
fail(
s"""
|Exception collecting dataset as objects
|${ds.exprEnc}
|${ds.exprEnc.deserializer.treeString}
|${ds.queryExecution}
""".stripMargin, e)
}
}
/**
* Runs the plan and makes sure the answer matches the expected result.
*
* @param df the [[DataFrame]] to be executed
* @param expectedAnswer the expected result in a [[Seq]] of [[Row]]s.
*/
protected def checkAnswer(df: => DataFrame, expectedAnswer: Seq[Row]): Unit = {
val analyzedDF = try df catch {
case ae: ExtendedAnalysisException =>
if (ae.plan.isDefined) {
fail(
s"""
|Failed to analyze query: $ae
|${ae.plan.get}
|
|${stackTraceToString(ae)}
|""".stripMargin)
} else {
throw ae
}
}
assertEmptyMissingInput(analyzedDF)
QueryTest.checkAnswer(analyzedDF, expectedAnswer)
}
protected def checkAnswer(df: => DataFrame, expectedAnswer: Row): Unit = {
checkAnswer(df, Seq(expectedAnswer))
}
protected def checkAnswer(df: => DataFrame, expectedAnswer: DataFrame): Unit = {
checkAnswer(df, expectedAnswer.collect().toImmutableArraySeq)
}
/**
* Runs the plan and makes sure the answer matches the expected result.
*
* @param df the [[DataFrame]] to be executed
* @param expectedAnswer the expected result in a [[Array]] of [[Row]]s.
*/
protected def checkAnswer(df: => DataFrame, expectedAnswer: Array[Row]): Unit = {
checkAnswer(df, expectedAnswer.toImmutableArraySeq)
}
/**
* Runs the plan and makes sure the answer is within absTol of the expected result.
*
* @param dataFrame the [[DataFrame]] to be executed
* @param expectedAnswer the expected result in a [[Seq]] of [[Row]]s.
* @param absTol the absolute tolerance between actual and expected answers.
*/
protected def checkAggregatesWithTol(dataFrame: DataFrame,
expectedAnswer: Seq[Row],
absTol: Double): Unit = {
// TODO: catch exceptions in data frame execution
val actualAnswer = dataFrame.collect()
require(actualAnswer.length == expectedAnswer.length,
s"actual num rows ${actualAnswer.length} != expected num of rows ${expectedAnswer.length}")
actualAnswer.zip(expectedAnswer).foreach {
case (actualRow, expectedRow) =>
QueryTest.checkAggregatesWithTol(actualRow, expectedRow, absTol)
}
}
protected def checkAggregatesWithTol(dataFrame: DataFrame,
expectedAnswer: Row,
absTol: Double): Unit = {
checkAggregatesWithTol(dataFrame, Seq(expectedAnswer), absTol)
}
/**
* Asserts that a given [[Dataset]] will be executed using the given number of cached results.
*/
def assertCached(query: Dataset[_], numCachedTables: Int = 1): Unit = {
val planWithCaching = query.queryExecution.withCachedData
val cachedData = planWithCaching collect {
case cached: InMemoryRelation => cached
}
assert(
cachedData.size == numCachedTables,
s"Expected query to contain $numCachedTables, but it actually had ${cachedData.size}\n" +
planWithCaching)
}
/**
* Asserts that a given [[Dataset]] will be executed using the cache with the given name and
* storage level.
*/
def assertCached(query: Dataset[_], cachedName: String, storageLevel: StorageLevel): Unit = {
val planWithCaching = query.queryExecution.withCachedData
val matched = planWithCaching.exists {
case cached: InMemoryRelation =>
val cacheBuilder = cached.cacheBuilder
cachedName == cacheBuilder.tableName.get && (storageLevel == cacheBuilder.storageLevel)
case _ => false
}
assert(matched, s"Expected query plan to hit cache $cachedName with storage " +
s"level $storageLevel, but it doesn't.")
}
/**
* Asserts that a given [[Dataset]] does not have missing inputs in all the analyzed plans.
*/
def assertEmptyMissingInput(query: Dataset[_]): Unit = {
assert(query.queryExecution.analyzed.missingInput.isEmpty,
s"The analyzed logical plan has missing inputs:\n${query.queryExecution.analyzed}")
assert(query.queryExecution.optimizedPlan.missingInput.isEmpty,
s"The optimized logical plan has missing inputs:\n${query.queryExecution.optimizedPlan}")
assert(query.queryExecution.executedPlan.missingInput.isEmpty,
s"The physical plan has missing inputs:\n${query.queryExecution.executedPlan}")
}
protected def getCurrentClassCallSitePattern: String = {
val cs = Thread.currentThread().getStackTrace()(2)
s"${cs.getClassName}\\..*\\(${cs.getFileName}:\\d+\\)"
}
protected def getNextLineCallSitePattern(lines: Int = 1): String = {
val cs = Thread.currentThread().getStackTrace()(2)
Pattern.quote(
s"${cs.getClassName}.${cs.getMethodName}(${cs.getFileName}:${cs.getLineNumber + lines})")
}
}
object QueryTest extends Assertions {
/**
* Runs the plan and makes sure the answer matches the expected result.
*
* @param df the DataFrame to be executed
* @param expectedAnswer the expected result in a Seq of Rows.
* @param checkToRDD whether to verify deserialization to an RDD. This runs the query twice.
*/
def checkAnswer(df: DataFrame, expectedAnswer: Seq[Row], checkToRDD: Boolean = true): Unit = {
getErrorMessageInCheckAnswer(df, expectedAnswer, checkToRDD) match {
case Some(errorMessage) => fail(errorMessage)
case None =>
}
}
/**
* Runs the plan and makes sure the answer matches the expected result.
* If there was exception during the execution or the contents of the DataFrame does not
* match the expected result, an error message will be returned. Otherwise, a None will
* be returned.
*
* @param df the DataFrame to be executed
* @param expectedAnswer the expected result in a Seq of Rows.
* @param checkToRDD whether to verify deserialization to an RDD. This runs the query twice.
*/
def getErrorMessageInCheckAnswer(
df: DataFrame,
expectedAnswer: Seq[Row],
checkToRDD: Boolean = true): Option[String] = {
val isSorted = df.logicalPlan.collect { case s: logical.Sort => s }.nonEmpty
if (checkToRDD) {
SQLExecution.withSQLConfPropagated(df.sparkSession) {
df.rdd.count() // Also attempt to deserialize as an RDD [SPARK-15791]
}
}
val sparkAnswer = try df.collect().toSeq catch {
case e: Exception =>
val errorMessage =
s"""
|Exception thrown while executing query:
|${df.queryExecution}
|== Exception ==
|$e
|${org.apache.spark.sql.catalyst.util.stackTraceToString(e)}
""".stripMargin
return Some(errorMessage)
}
sameRows(expectedAnswer, sparkAnswer, isSorted).map { results =>
s"""
|Results do not match for query:
|Timezone: ${TimeZone.getDefault}
|Timezone Env: ${sys.env.getOrElse("TZ", "")}
|
|${df.queryExecution}
|== Results ==
|$results
""".stripMargin
}
}
def prepareAnswer(answer: Seq[Row], isSorted: Boolean): Seq[Row] = {
// Converts data to types that we can do equality comparison using Scala collections.
// For BigDecimal type, the Scala type has a better definition of equality test (similar to
// Java's java.math.BigDecimal.compareTo).
// For binary arrays, we convert it to Seq to avoid of calling java.util.Arrays.equals for
// equality test.
val converted: Seq[Row] = answer.map(prepareRow)
if (!isSorted) converted.sortBy(_.toString()) else converted
}
// We need to call prepareRow recursively to handle schemas with struct types.
def prepareRow(row: Row): Row = {
Row.fromSeq(row.toSeq.map {
case null => null
case bd: java.math.BigDecimal => BigDecimal(bd)
// Equality of WrappedArray differs for AnyVal and AnyRef in Scala 2.12.2+
case seq: Seq[_] => seq.map {
case b: java.lang.Byte => b.byteValue
case s: java.lang.Short => s.shortValue
case i: java.lang.Integer => i.intValue
case l: java.lang.Long => l.longValue
case f: java.lang.Float => f.floatValue
case d: java.lang.Double => d.doubleValue
case x => x
}
// Convert array to Seq for easy equality check.
case b: Array[_] => b.toSeq
case r: Row => prepareRow(r)
case o => o
})
}
private def genError(
expectedAnswer: Seq[Row],
sparkAnswer: Seq[Row],
isSorted: Boolean = false): String = {
val getRowType: Option[Row] => String = row =>
row.map(row =>
if (row.schema == null) {
"struct<>"
} else {
s"${row.schema.catalogString}"
}).getOrElse("struct<>")
s"""
|== Results ==
|${
sideBySide(
s"== Correct Answer - ${expectedAnswer.size} ==" +:
getRowType(expectedAnswer.headOption) +:
prepareAnswer(expectedAnswer, isSorted).map(_.toString()),
s"== Spark Answer - ${sparkAnswer.size} ==" +:
getRowType(sparkAnswer.headOption) +:
prepareAnswer(sparkAnswer, isSorted).map(_.toString())).mkString("\n")
}
""".stripMargin
}
def includesRows(
expectedRows: Seq[Row],
sparkAnswer: Seq[Row]): Option[String] = {
if (!prepareAnswer(expectedRows, true).toSet.subsetOf(prepareAnswer(sparkAnswer, true).toSet)) {
return Some(genError(expectedRows, sparkAnswer, true))
}
None
}
def compare(obj1: Any, obj2: Any): Boolean = (obj1, obj2) match {
case (null, null) => true
case (null, _) => false
case (_, null) => false
case (a: Array[_], b: Array[_]) =>
a.length == b.length && a.zip(b).forall { case (l, r) => compare(l, r)}
case (a: Map[_, _], b: Map[_, _]) =>
a.size == b.size && a.keys.forall { aKey =>
b.keys.find(bKey => compare(aKey, bKey)).exists(bKey => compare(a(aKey), b(bKey)))
}
case (a: Iterable[_], b: Iterable[_]) =>
a.size == b.size && a.zip(b).forall { case (l, r) => compare(l, r)}
case (a: Product, b: Product) =>
compare(a.productIterator.toSeq, b.productIterator.toSeq)
case (a: Row, b: Row) =>
compare(a.toSeq, b.toSeq)
// 0.0 == -0.0, turn float/double to bits before comparison, to distinguish 0.0 and -0.0.
case (a: Double, b: Double) =>
java.lang.Double.doubleToRawLongBits(a) == java.lang.Double.doubleToRawLongBits(b)
case (a: Float, b: Float) =>
java.lang.Float.floatToRawIntBits(a) == java.lang.Float.floatToRawIntBits(b)
case (a, b) => a == b
}
def sameRows(
expectedAnswer: Seq[Row],
sparkAnswer: Seq[Row],
isSorted: Boolean = false): Option[String] = {
if (!compare(prepareAnswer(expectedAnswer, isSorted), prepareAnswer(sparkAnswer, isSorted))) {
return Some(genError(expectedAnswer, sparkAnswer, isSorted))
}
None
}
/**
* Runs the plan and makes sure the answer is within absTol of the expected result.
*
* @param actualAnswer the actual result in a [[Row]].
* @param expectedAnswer the expected result in a[[Row]].
* @param absTol the absolute tolerance between actual and expected answers.
*/
protected def checkAggregatesWithTol(actualAnswer: Row, expectedAnswer: Row, absTol: Double) = {
require(actualAnswer.length == expectedAnswer.length,
s"actual answer length ${actualAnswer.length} != " +
s"expected answer length ${expectedAnswer.length}")
// TODO: support other numeric types besides Double
// TODO: support struct types?
actualAnswer.toSeq.zip(expectedAnswer.toSeq).foreach {
case (actual: Double, expected: Double) =>
assert(math.abs(actual - expected) < absTol,
s"actual answer $actual not within $absTol of correct answer $expected")
case (actual, expected) =>
assert(actual == expected, s"$actual did not equal $expected")
}
}
def checkAnswer(df: DataFrame, expectedAnswer: java.util.List[Row]): Unit = {
getErrorMessageInCheckAnswer(df, expectedAnswer.asScala.toSeq) match {
case Some(errorMessage) => fail(errorMessage)
case None =>
}
}
}
class QueryTestSuite extends QueryTest with test.SharedSparkSession {
test("SPARK-16940: checkAnswer should raise TestFailedException for wrong results") {
intercept[org.scalatest.exceptions.TestFailedException] {
checkAnswer(sql("SELECT 1"), Row(2) :: Nil)
}
}
}