ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

[SPARK-13651] Generator outputs are not resolved correctly resulting in run time error

Browse source 
## What changes were proposed in this pull request?

```
Seq(("id1", "value1")).toDF("key", "value").registerTempTable("src")
sqlContext.sql("SELECT t1.* FROM src LATERAL VIEW explode(map('key1', 100, 'key2', 200)) t1 AS key, value")
```
Results in following logical plan

```
Project [key#2,value#3]
+- Generate explode(HiveGenericUDF#org.apache.hadoop.hive.ql.udf.generic.GenericUDFMap(key1,100,key2,200)), true, false, Some(genoutput), [key#2,value#3]
   +- SubqueryAlias src
      +- Project [_1#0 AS key#2,_2#1 AS value#3]
         +- LocalRelation [_1#0,_2#1], [[id1,value1]]
```

The above query fails with following runtime error.
```
java.lang.ClassCastException: java.lang.Integer cannot be cast to org.apache.spark.unsafe.types.UTF8String
	at org.apache.spark.sql.catalyst.expressions.BaseGenericInternalRow$class.getUTF8String(rows.scala:46)
	at org.apache.spark.sql.catalyst.expressions.GenericInternalRow.getUTF8String(rows.scala:221)
	at org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.apply(generated.java:42)
	at org.apache.spark.sql.execution.Generate$$anonfun$doExecute$1$$anonfun$apply$9.apply(Generate.scala:98)
	at org.apache.spark.sql.execution.Generate$$anonfun$doExecute$1$$anonfun$apply$9.apply(Generate.scala:96)
	at scala.collection.Iterator$$anon$11.next(Iterator.scala:370)
	at scala.collection.Iterator$$anon$11.next(Iterator.scala:370)
	at scala.collection.Iterator$class.foreach(Iterator.scala:742)
	at scala.collection.AbstractIterator.foreach(Iterator.scala:1194)
        <stack-trace omitted.....>
```
In this case the generated outputs are wrongly resolved from its child (LocalRelation) due to
https://github.com/apache/spark/blob/master/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala#L537-L548
## How was this patch tested?

(Please explain how this patch was tested. E.g. unit tests, integration tests, manual tests)

Added unit tests in hive/SQLQuerySuite and AnalysisSuite

Author: Dilip Biswal <dbiswal@us.ibm.com>

Closes #11497 from dilipbiswal/spark-13651.
This commit is contained in:
Dilip Biswal 2016-03-07 09:46:28 -08:00 committed by Davies Liu
parent 03f57a6c2d
commit d7eac9d795
2 changed files with 13 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

5
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/Analyzer.scala
View file

@ -512,8 +512,9 @@ class Analyzer(
// A special case for Generate, because the output of Generate should not be resolved by
// ResolveReferences. Attributes in the output will be resolved by ResolveGenerate.
case g @ Generate(generator, join, outer, qualifier, output, child)
if child.resolved && !generator.resolved =>
case g @ Generate(generator, _, _, _, _, _) if generator.resolved => g
case g @ Generate(generator, join, outer, qualifier, output, child) =>
val newG = resolveExpression(generator, child, throws = true)
if (newG.fastEquals(generator)) {
g

10
sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/SQLQuerySuite.scala
View file

@ -92,6 +92,16 @@ class SQLQuerySuite extends QueryTest with SQLTestUtils with TestHiveSingleton {
checkAnswer(query, Row(1, 1) :: Row(1, 2) :: Row(1, 3) :: Nil)
}
test("SPARK-13651: generator outputs shouldn't be resolved from its child's output") {
withTempTable("src") {
Seq(("id1", "value1")).toDF("key", "value").registerTempTable("src")
val query =
sql("SELECT genoutput.* FROM src " +
"LATERAL VIEW explode(map('key1', 100, 'key2', 200)) genoutput AS key, value")
checkAnswer(query, Row("key1", 100) :: Row("key2", 200) :: Nil)
}
}
test("SPARK-6851: Self-joined converted parquet tables") {
val orders = Seq(
Order(1, "Atlas", "MTB", 234, "2015-01-07", "John D", "Pacifica", "CA", 20151),