[SPARK-15192] RowEncoder needs to verify nullability in a more explicit way

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 2.0.0
• Component/s: SQL
• Target Version/s: 2.0.0
• Assignee: Wenchen Fan
• Reporter: Yin Huai
• Created: 06/May/16 21:36
• Updated: 19/May/16 01:08
• Resolved: 19/May/16 01:08

Description

When we create a Dataset from an RDD of rows with a specific schema, if the nullability of a value does not match the nullability defined in the schema, we will throw an exception that is not easy to understand.
It will be good to verify the nullability in a more explicit way.

import org.apache.spark.sql.types._
import org.apache.spark.sql.Row

val schema = new StructType().add("a", StringType, false).add("b", StringType, false)
val rdd = sc.parallelize(Row(null, "123") :: Row("234", null) :: Nil)
spark.createDataFrame(rdd, schema).show

java.lang.RuntimeException: Error while decoding: java.lang.NullPointerException
createexternalrow(if (isnull(input[0, string])) null else input[0, string].toString, if (isnull(input[1, string])) null else input[1, string].toString, StructField(a,StringType,false), StructField(b,StringType,false))
:- if (isnull(input[0, string])) null else input[0, string].toString
:  :- isnull(input[0, string])
:  :  +- input[0, string]
:  :- null
:  +- input[0, string].toString
:     +- input[0, string]
+- if (isnull(input[1, string])) null else input[1, string].toString
   :- isnull(input[1, string])
   :  +- input[1, string]
   :- null
   +- input[1, string].toString
      +- input[1, string]

  at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder.fromRow(ExpressionEncoder.scala:244)
  at org.apache.spark.sql.Dataset$$anonfun$org$apache$spark$sql$Dataset$$execute$1$1$$anonfun$apply$13.apply(Dataset.scala:2119)
  at org.apache.spark.sql.Dataset$$anonfun$org$apache$spark$sql$Dataset$$execute$1$1$$anonfun$apply$13.apply(Dataset.scala:2119)
  at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:234)
  at scala.collection.TraversableLike$$anonfun$map$1.apply(TraversableLike.scala:234)
  at scala.collection.IndexedSeqOptimized$class.foreach(IndexedSeqOptimized.scala:33)
  at scala.collection.mutable.ArrayOps$ofRef.foreach(ArrayOps.scala:186)
  at scala.collection.TraversableLike$class.map(TraversableLike.scala:234)
  at scala.collection.mutable.ArrayOps$ofRef.map(ArrayOps.scala:186)
  at org.apache.spark.sql.Dataset$$anonfun$org$apache$spark$sql$Dataset$$execute$1$1.apply(Dataset.scala:2119)
  at org.apache.spark.sql.execution.SQLExecution$.withNewExecutionId(SQLExecution.scala:57)
  at org.apache.spark.sql.Dataset.withNewExecutionId(Dataset.scala:2407)
  at org.apache.spark.sql.Dataset.org$apache$spark$sql$Dataset$$execute$1(Dataset.scala:2118)
  at org.apache.spark.sql.Dataset.org$apache$spark$sql$Dataset$$collect(Dataset.scala:2125)
  at org.apache.spark.sql.Dataset$$anonfun$head$1.apply(Dataset.scala:1859)
  at org.apache.spark.sql.Dataset$$anonfun$head$1.apply(Dataset.scala:1858)
  at org.apache.spark.sql.Dataset.withTypedCallback(Dataset.scala:2437)
  at org.apache.spark.sql.Dataset.head(Dataset.scala:1858)
  at org.apache.spark.sql.Dataset.take(Dataset.scala:2075)
  at org.apache.spark.sql.Dataset.showString(Dataset.scala:239)
  at org.apache.spark.sql.Dataset.show(Dataset.scala:530)
  at org.apache.spark.sql.Dataset.show(Dataset.scala:490)
  at org.apache.spark.sql.Dataset.show(Dataset.scala:499)
  ... 50 elided
Caused by: java.lang.NullPointerException
  at org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificSafeProjection.apply(Unknown Source)
  at org.apache.spark.sql.catalyst.encoders.ExpressionEncoder.fromRow(ExpressionEncoder.scala:241)
  ... 72 more

Issue Links

• relates to: SPARK-13740 add null check for _verify_type in types.py
• links to: [Github] Pull Request #13008 (cloud-fan)