[SPARK-11188][SQL] Elide stacktraces in bin/spark-sql for AnalysisExceptions

Only print the error message to the console for Analysis Exceptions in sql-shell.

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

Closes #9194 from dilipbiswal/spark-11188.

3 files changed, 27 insertions, 6 deletions

diff --git a/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLCLIDriver.scala b/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLCLIDriver.scala
index b5073961a1..62e912c69a 100644
--- a/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLCLIDriver.scala
+++ b/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLCLIDriver.scala
@@ -20,6 +20,8 @@ package org.apache.spark.sql.hive.thriftserver
 import java.io._
 import java.util.{ArrayList => JArrayList, Locale}
 
+import org.apache.spark.sql.AnalysisException
+
 import scala.collection.JavaConverters._
 
 import jline.console.ConsoleReader
@@ -298,6 +300,7 @@ private[hive] class SparkSQLCLIDriver extends CliDriver with Logging {
 
           driver.init()
           val out = sessionState.out
+          val err = sessionState.err
           val start: Long = System.currentTimeMillis()
           if (sessionState.getIsVerbose) {
             out.println(cmd)
@@ -308,7 +311,12 @@ private[hive] class SparkSQLCLIDriver extends CliDriver with Logging {
 
           ret = rc.getResponseCode
           if (ret != 0) {
-            console.printError(rc.getErrorMessage())
+            // For analysis exception, only the error is printed out to the console.
+            rc.getException() match {
+              case e : AnalysisException =>
+                err.println(s"""Error in query: ${e.getMessage}""")
+              case _ => err.println(rc.getErrorMessage())
+            }
             driver.close()
             return ret
           }
diff --git a/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLDriver.scala b/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLDriver.scala
index 2619286afc..f1ec723852 100644
--- a/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLDriver.scala
+++ b/sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/SparkSQLDriver.scala
@@ -18,6 +18,8 @@
 package org.apache.spark.sql.hive.thriftserver
 
 import java.util.{Arrays, ArrayList => JArrayList, List => JList}
+import org.apache.log4j.LogManager
+import org.apache.spark.sql.AnalysisException
 
 import scala.collection.JavaConverters._
 
@@ -63,9 +65,12 @@ private[hive] class SparkSQLDriver(
       tableSchema = getResultSetSchema(execution)
       new CommandProcessorResponse(0)
     } catch {
-      case cause: Throwable =>
-        logError(s"Failed in [$command]", cause)
-        new CommandProcessorResponse(1, ExceptionUtils.getStackTrace(cause), null)
+        case ae: AnalysisException =>
+          logDebug(s"Failed in [$command]", ae)
+          new CommandProcessorResponse(1, ExceptionUtils.getStackTrace(ae), null, ae)
+        case cause: Throwable =>
+          logError(s"Failed in [$command]", cause)
+          new CommandProcessorResponse(1, ExceptionUtils.getStackTrace(cause), null, cause)
     }
   }
 
diff --git a/sql/hive-thriftserver/src/test/scala/org/apache/spark/sql/hive/thriftserver/CliSuite.scala b/sql/hive-thriftserver/src/test/scala/org/apache/spark/sql/hive/thriftserver/CliSuite.scala
index 76d1591a23..3fa5c8528b 100644
--- a/sql/hive-thriftserver/src/test/scala/org/apache/spark/sql/hive/thriftserver/CliSuite.scala
+++ b/sql/hive-thriftserver/src/test/scala/org/apache/spark/sql/hive/thriftserver/CliSuite.scala
@@ -58,7 +58,7 @@ class CliSuite extends SparkFunSuite with BeforeAndAfter with Logging {
    * @param timeout maximum time for the commands to complete
    * @param extraArgs any extra arguments
    * @param errorResponses a sequence of strings whose presence in the stdout of the forked process
-   *                       is taken as an immediate error condition. That is: if a line beginning
+   *                       is taken as an immediate error condition. That is: if a line containing
    *                       with one of these strings is found, fail the test immediately.
    *                       The default value is `Seq("Error:")`
    *
@@ -104,7 +104,7 @@ class CliSuite extends SparkFunSuite with BeforeAndAfter with Logging {
         }
       } else {
         errorResponses.foreach { r =>
-          if (line.startsWith(r)) {
+          if (line.contains(r)) {
             foundAllExpectedAnswers.tryFailure(
               new RuntimeException(s"Failed with error line '$line'"))
           }
@@ -219,4 +219,12 @@ class CliSuite extends SparkFunSuite with BeforeAndAfter with Logging {
         -> "OK"
     )
   }
+
+  test("SPARK-11188 Analysis error reporting") {
+    runCliWithin(timeout = 2.minute,
+      errorResponses = Seq("AnalysisException"))(
+      "select * from nonexistent_table;"
+        -> "Error in query: Table not found: nonexistent_table;"
+    )
+  }
 }