[SPARK-17011][SQL] Support testing exceptions in SQLQueryTestSuite

## What changes were proposed in this pull request?
This patch adds exception testing to SQLQueryTestSuite. When there is an exception in query execution, the query result contains the the exception class along with the exception message.

As part of this, I moved some additional test cases for limit from SQLQuerySuite over to SQLQueryTestSuite.

## How was this patch tested?
This is a test harness change.

Author: petermaxlee <petermaxlee@gmail.com>

Closes #14592 from petermaxlee/SPARK-17011.

Author: petermaxlee

sql/core/src/test/resources/sql-tests/inputs/limit.sql

@@ -0,0 +1,20 @@
+
+-- limit on various data types
+select * from testdata limit 2;
+select * from arraydata limit 2;
+select * from mapdata limit 2;
+
+-- foldable non-literal in limit
+select * from testdata limit 2 + 1;
+
+select * from testdata limit CAST(1 AS int);
+
+-- limit must be non-negative
+select * from testdata limit -1;
+
+-- limit must be foldable
+select * from testdata limit key > 3;
+
+-- limit must be integer
+select * from testdata limit true;
+select * from testdata limit 'a';

sql/core/src/test/resources/sql-tests/inputs/number-format.sql

@@ -3,10 +3,13 @@
 -- parse as ints
 select 1, -1;
 
--- parse as longs
+-- parse as longs (Int.MaxValue + 1, and Int.MinValue - 1)
 select 2147483648, -2147483649;
 
--- parse as decimals
+-- parse long min and max value
+select 9223372036854775807, -9223372036854775808;
+
+-- parse as decimals (Long.MaxValue + 1, and Long.MinValue - 1)
 select 9223372036854775808, -9223372036854775809;
 
 -- various floating point (decimal) formats

sql/core/src/test/resources/sql-tests/results/datetime.sql.out

@@ -1,4 +1,4 @@
--- Automatically generated by org.apache.spark.sql.SQLQueryTestSuite
+-- Automatically generated by SQLQueryTestSuite
 -- Number of queries: 1
 
 

sql/core/src/test/resources/sql-tests/results/having.sql.out

@@ -1,4 +1,4 @@
--- Automatically generated by org.apache.spark.sql.SQLQueryTestSuite
+-- Automatically generated by SQLQueryTestSuite
 -- Number of queries: 4
 
 

sql/core/src/test/resources/sql-tests/results/limit.sql.out

@@ -0,0 +1,83 @@
+-- Automatically generated by SQLQueryTestSuite
+-- Number of queries: 9
+
+
+-- !query 0
+select * from testdata limit 2
+-- !query 0 schema
+struct<key:int,value:string>
+-- !query 0 output
+1	1
+2	2
+
+
+-- !query 1
+select * from arraydata limit 2
+-- !query 1 schema
+struct<arraycol:array<int>,nestedarraycol:array<array<int>>>
+-- !query 1 output
+[1,2,3]	[[1,2,3]]
+[2,3,4]	[[2,3,4]]
+
+
+-- !query 2
+select * from mapdata limit 2
+-- !query 2 schema
+struct<mapcol:map<int,string>>
+-- !query 2 output
+{1:"a1",2:"b1",3:"c1",4:"d1",5:"e1"}
+{1:"a2",2:"b2",3:"c2",4:"d2"}
+
+
+-- !query 3
+select * from testdata limit 2 + 1
+-- !query 3 schema
+struct<key:int,value:string>
+-- !query 3 output
+1	1
+2	2
+3	3
+
+
+-- !query 4
+select * from testdata limit CAST(1 AS int)
+-- !query 4 schema
+struct<key:int,value:string>
+-- !query 4 output
+1	1
+
+
+-- !query 5
+select * from testdata limit -1
+-- !query 5 schema
+struct<>
+-- !query 5 output
+org.apache.spark.sql.AnalysisException
+The limit expression must be equal to or greater than 0, but got -1;
+
+
+-- !query 6
+select * from testdata limit key > 3
+-- !query 6 schema
+struct<>
+-- !query 6 output
+org.apache.spark.sql.AnalysisException
+The limit expression must evaluate to a constant value, but got (testdata.`key` > 3);
+
+
+-- !query 7
+select * from testdata limit true
+-- !query 7 schema
+struct<>
+-- !query 7 output
+org.apache.spark.sql.AnalysisException
+The limit expression must be integer type, but got boolean;
+
+
+-- !query 8
+select * from testdata limit 'a'
+-- !query 8 schema
+struct<>
+-- !query 8 output
+org.apache.spark.sql.AnalysisException
+The limit expression must be integer type, but got string;

sql/core/src/test/resources/sql-tests/results/natural-join.sql.out

@@ -1,4 +1,4 @@
--- Automatically generated by org.apache.spark.sql.SQLQueryTestSuite
+-- Automatically generated by SQLQueryTestSuite
 -- Number of queries: 6
 
 

sql/core/src/test/resources/sql-tests/results/number-format.sql.out

@@ -1,5 +1,5 @@
--- Automatically generated by org.apache.spark.sql.SQLQueryTestSuite
--- Number of queries: 4
+-- Automatically generated by SQLQueryTestSuite
+-- Number of queries: 5
 
 
 -- !query 0
@@ -19,16 +19,24 @@ struct<2147483648:bigint,(-2147483649):bigint>
 
 
 -- !query 2
-select 9223372036854775808, -9223372036854775809
+select 9223372036854775807, -9223372036854775808
 -- !query 2 schema
-struct<9223372036854775808:decimal(19,0),(-9223372036854775809):decimal(19,0)>
+struct<9223372036854775807:bigint,(-9223372036854775808):decimal(19,0)>
 -- !query 2 output
-9223372036854775808	-9223372036854775809
+9223372036854775807	-9223372036854775808
 
 
 -- !query 3
-select 0.3, -0.8, .5, -.18, 0.1111
+select 9223372036854775808, -9223372036854775809
 -- !query 3 schema
-struct<0.3:decimal(1,1),(-0.8):decimal(1,1),0.5:decimal(1,1),(-0.18):decimal(2,2),0.1111:decimal(4,4)>
+struct<9223372036854775808:decimal(19,0),(-9223372036854775809):decimal(19,0)>
 -- !query 3 output
+9223372036854775808	-9223372036854775809
+
+
+-- !query 4
+select 0.3, -0.8, .5, -.18, 0.1111
+-- !query 4 schema
+struct<0.3:decimal(1,1),(-0.8):decimal(1,1),0.5:decimal(1,1),(-0.18):decimal(2,2),0.1111:decimal(4,4)>
+-- !query 4 output
 0.3	-0.8	0.5	-0.18	0.1111

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

@@ -650,51 +650,12 @@ class SQLQuerySuite extends QueryTest with SharedSQLContext {
     sortTest()
   }
 
-  test("limit") {
-    checkAnswer(
-      sql("SELECT * FROM testData LIMIT 9 + 1"),
-      testData.take(10).toSeq)
-
-    checkAnswer(
-      sql("SELECT * FROM arrayData LIMIT CAST(1 AS Integer)"),
-      arrayData.collect().take(1).map(Row.fromTuple).toSeq)
-
-    checkAnswer(
-      sql("SELECT * FROM mapData LIMIT 1"),
-      mapData.collect().take(1).map(Row.fromTuple).toSeq)
-  }
-
-  test("non-foldable expressions in LIMIT") {
-    val e = intercept[AnalysisException] {
-      sql("SELECT * FROM testData LIMIT key > 3")
-    }.getMessage
-    assert(e.contains("The limit expression must evaluate to a constant value, " +
-      "but got (testdata.`key` > 3)"))
-  }
-
-  test("Expressions in limit clause are not integer") {
-    var e = intercept[AnalysisException] {
-      sql("SELECT * FROM testData LIMIT true")
-    }.getMessage
-    assert(e.contains("The limit expression must be integer type, but got boolean"))
-
-    e = intercept[AnalysisException] {
-      sql("SELECT * FROM testData LIMIT 'a'")
-    }.getMessage
-    assert(e.contains("The limit expression must be integer type, but got string"))
-  }
-
   test("negative in LIMIT or TABLESAMPLE") {
     val expected = "The limit expression must be equal to or greater than 0, but got -1"
     var e = intercept[AnalysisException] {
       sql("SELECT * FROM testData TABLESAMPLE (-1 rows)")
     }.getMessage
     assert(e.contains(expected))
-
-    e = intercept[AnalysisException] {
-      sql("SELECT * FROM testData LIMIT -1")
-    }.getMessage
-    assert(e.contains(expected))
   }
 
   test("CTE feature") {
@@ -1337,17 +1298,6 @@ class SQLQuerySuite extends QueryTest with SharedSQLContext {
     }
   }
 
-  test("Test to check we can use Long.MinValue") {
-    checkAnswer(
-      sql(s"SELECT ${Long.MinValue} FROM testData ORDER BY key LIMIT 1"), Row(Long.MinValue)
-    )
-
-    checkAnswer(
-      sql(s"SELECT key FROM testData WHERE key > ${Long.MinValue}"),
-      (1 to 100).map(Row(_)).toSeq
-    )
-  }
-
   test("Test to check we can apply sign to expression") {
 
     checkAnswer(

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

@@ -20,6 +20,8 @@ package org.apache.spark.sql
 import java.io.File
 import java.util.{Locale, TimeZone}
 
+import scala.util.control.NonFatal
+
 import org.apache.spark.sql.catalyst.planning.PhysicalOperation
 import org.apache.spark.sql.catalyst.plans.logical._
 import org.apache.spark.sql.catalyst.rules.RuleExecutor
@@ -132,6 +134,7 @@ class SQLQueryTestSuite extends QueryTest with SharedSQLContext {
     // Create a local SparkSession to have stronger isolation between different test cases.
     // This does not isolate catalog changes.
     val localSparkSession = spark.newSession()
+    loadTestData(localSparkSession)
 
     // Run the SQL queries preparing them for comparison.
     val outputs: Seq[QueryOutput] = queries.map { sql =>
@@ -146,7 +149,7 @@ class SQLQueryTestSuite extends QueryTest with SharedSQLContext {
     if (regenerateGoldenFiles) {
       // Again, we are explicitly not using multi-line string due to stripMargin removing "|".
       val goldenOutput = {
-        s"-- Automatically generated by ${getClass.getName}\n" +
+        s"-- Automatically generated by ${getClass.getSimpleName}\n" +
         s"-- Number of queries: ${outputs.size}\n\n\n" +
         outputs.zipWithIndex.map{case (qr, i) => qr.toString(i)}.mkString("\n\n\n") + "\n"
       }
@@ -192,12 +195,19 @@ class SQLQueryTestSuite extends QueryTest with SharedSQLContext {
       case _ => plan.children.iterator.exists(isSorted)
     }
 
-    val df = session.sql(sql)
-    val schema = df.schema
-    val answer = df.queryExecution.hiveResultString()
+    try {
+      val df = session.sql(sql)
+      val schema = df.schema
+      val answer = df.queryExecution.hiveResultString()
+
+      // If the output is not pre-sorted, sort it.
+      if (isSorted(df.queryExecution.analyzed)) (schema, answer) else (schema, answer.sorted)
 
-    // If the output is not pre-sorted, sort it.
-    if (isSorted(df.queryExecution.analyzed)) (schema, answer) else (schema, answer.sorted)
+    } catch {
+      case NonFatal(e) =>
+        // If there is an exception, put the exception class followed by the message.
+        (StructType(Seq.empty), Seq(e.getClass.getName, e.getMessage))
+    }
   }
 
   private def listTestCases(): Seq[TestCase] = {
@@ -213,6 +223,25 @@ class SQLQueryTestSuite extends QueryTest with SharedSQLContext {
     files ++ dirs.flatMap(listFilesRecursively)
   }
 
+  /** Load built-in test tables into the SparkSession. */
+  private def loadTestData(session: SparkSession): Unit = {
+    import session.implicits._
+
+    (1 to 100).map(i => (i, i.toString)).toDF("key", "value").createOrReplaceTempView("testdata")
+
+    ((Seq(1, 2, 3), Seq(Seq(1, 2, 3))) :: (Seq(2, 3, 4), Seq(Seq(2, 3, 4))) :: Nil)
+      .toDF("arraycol", "nestedarraycol")
+      .createOrReplaceTempView("arraydata")
+
+    (Tuple1(Map(1 -> "a1", 2 -> "b1", 3 -> "c1", 4 -> "d1", 5 -> "e1")) ::
+      Tuple1(Map(1 -> "a2", 2 -> "b2", 3 -> "c2", 4 -> "d2")) ::
+      Tuple1(Map(1 -> "a3", 2 -> "b3", 3 -> "c3")) ::
+      Tuple1(Map(1 -> "a4", 2 -> "b4")) ::
+      Tuple1(Map(1 -> "a5")) :: Nil)
+      .toDF("mapcol")
+      .createOrReplaceTempView("mapdata")
+  }
+
   private val originalTimeZone = TimeZone.getDefault
   private val originalLocale = Locale.getDefault