[SPARK-21954][SQL] JacksonUtils should verify MapType's value type in…

…stead of key type

## What changes were proposed in this pull request?

`JacksonUtils.verifySchema` verifies if a data type can be converted to JSON. For `MapType`, it now verifies the key type. However, in `JacksonGenerator`, when converting a map to JSON, we only care about its values and create a writer for the values. The keys in a map are treated as strings by calling `toString` on the keys.

Thus, we should change `JacksonUtils.verifySchema` to verify the value type of `MapType`.

## How was this patch tested?

Added tests.

Author: Liang-Chi Hsieh <viirya@gmail.com>

Closes apache#19167 from viirya/test-jacksonutils.

sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/json/JacksonUtils.scala

@@ -44,7 +44,9 @@ object JacksonUtils {
 
       case at: ArrayType => verifyType(name, at.elementType)
 
-      case mt: MapType => verifyType(name, mt.keyType)
+      // For MapType, its keys are treated as a string (i.e. calling `toString`) basically when
+      // generating JSON, so we only care if the values are valid for JSON.
+      case mt: MapType => verifyType(name, mt.valueType)
 
       case udt: UserDefinedType[_] => verifyType(name, udt.sqlType)
 

sql/catalyst/src/test/scala/org/apache/spark/sql/catalyst/expressions/JsonExpressionsSuite.scala

@@ -21,6 +21,7 @@ import java.util.Calendar
 
 import org.apache.spark.SparkFunSuite
 import org.apache.spark.sql.catalyst.InternalRow
+import org.apache.spark.sql.catalyst.errors.TreeNodeException
 import org.apache.spark.sql.catalyst.util.{DateTimeTestUtils, DateTimeUtils, GenericArrayData, PermissiveMode}
 import org.apache.spark.sql.types._
 import org.apache.spark.unsafe.types.UTF8String
@@ -610,4 +611,26 @@ class JsonExpressionsSuite extends SparkFunSuite with ExpressionEvalHelper {
       """{"t":"2015-12-31T16:00:00"}"""
     )
   }
+
+  test("to_json: verify MapType's value type instead of key type") {
+    // Keys in map are treated as strings when converting to JSON. The type doesn't matter at all.
+    val mapType1 = MapType(CalendarIntervalType, IntegerType)
+    val schema1 = StructType(StructField("a", mapType1) :: Nil)
+    val struct1 = Literal.create(null, schema1)
+    checkEvaluation(
+      StructsToJson(Map.empty, struct1, gmtId),
+      null
+    )
+
+    // The value type must be valid for converting to JSON.
+    val mapType2 = MapType(IntegerType, CalendarIntervalType)
+    val schema2 = StructType(StructField("a", mapType2) :: Nil)
+    val struct2 = Literal.create(null, schema2)
+    intercept[TreeNodeException[_]] {
+      checkEvaluation(
+        StructsToJson(Map.empty, struct2, gmtId),
+        null
+      )
+    }
+  }
 }

sql/core/src/test/scala/org/apache/spark/sql/JsonFunctionsSuite.scala

@@ -17,7 +17,7 @@
 
 package org.apache.spark.sql
 
-import org.apache.spark.sql.functions.{from_json, struct, to_json}
+import org.apache.spark.sql.functions.{from_json, lit, map, struct, to_json}
 import org.apache.spark.sql.test.SharedSQLContext
 import org.apache.spark.sql.types._
 
@@ -195,15 +195,33 @@ class JsonFunctionsSuite extends QueryTest with SharedSQLContext {
       Row("""{"_1":"26/08/2015 18:00"}""") :: Nil)
   }
 
-  test("to_json unsupported type") {
+  test("to_json - key types of map don't matter") {
+    // interval type is invalid for converting to JSON. However, the keys of a map are treated
+    // as strings, so its type doesn't matter.
     val df = Seq(Tuple1(Tuple1("interval -3 month 7 hours"))).toDF("a")
-      .select(struct($"a._1".cast(CalendarIntervalType).as("a")).as("c"))
+      .select(struct(map($"a._1".cast(CalendarIntervalType), lit("a")).as("col1")).as("c"))
+    checkAnswer(
+      df.select(to_json($"c")),
+      Row("""{"col1":{"interval -3 months 7 hours":"a"}}""") :: Nil)
+  }
+
+  test("to_json unsupported type") {
+    val baseDf = Seq(Tuple1(Tuple1("interval -3 month 7 hours"))).toDF("a")
+    val df = baseDf.select(struct($"a._1".cast(CalendarIntervalType).as("a")).as("c"))
     val e = intercept[AnalysisException]{
       // Unsupported type throws an exception
       df.select(to_json($"c")).collect()
     }
     assert(e.getMessage.contains(
       "Unable to convert column a of type calendarinterval to JSON."))
+
+    // interval type is invalid for converting to JSON. We can't use it as value type of a map.
+    val df2 = baseDf
+      .select(struct(map(lit("a"), $"a._1".cast(CalendarIntervalType)).as("col1")).as("c"))
+    val e2 = intercept[AnalysisException] {
+      df2.select(to_json($"c")).collect()
+    }
+    assert(e2.getMessage.contains("Unable to convert column col1 of type calendarinterval to JSON"))
   }
 
   test("roundtrip in to_json and from_json - struct") {