TPCH Benchmark fixes (#44)

dotnet · Apr 24, 2019 · 8719e44 · 8719e44
1 parent 43775f1
commit 8719e44
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 3 deletions.
diff --git a/benchmark/csharp/Tpch/Program.cs b/benchmark/csharp/Tpch/Program.cs
@@ -18,6 +18,8 @@ private static void Main(string[] args)
                 Console.WriteLine("\t<spark-submit> --master local");
                 Console.WriteLine("\t\t--class org.apache.spark.deploy.DotnetRunner <path-to-microsoft-spark-jar>");
                 Console.WriteLine("\t\tTpch.exe <tpch_data_root_path> <query_number> <num_iterations> <true for SQL | false for functional>");
+
+                return;
             }
 
             var tpchRoot = args[0];

diff --git a/benchmark/scala/src/main/scala/com/microsoft/tpch/App.scala b/benchmark/scala/src/main/scala/com/microsoft/tpch/App.scala
@@ -14,8 +14,10 @@ object App {
   def main(args: Array[String]) {
     if (args.length != 4) {
       println("Usage:")
-      println("\t<spark-submit> --master local --class com.microsoft.tpch.App microsoft-spark-examples-<version>.jar")
+      println("\t<spark-submit> --master local --class com.microsoft.tpch.App microsoft-spark-benchmark-<version>.jar")
       println("\t\t<tpch_data_root_path> <query_number> <num_iterations> <true for SQL | false for functional>")
+
+      return
     }
 
     val tpchRoot = args(0)

diff --git a/benchmark/scala/src/main/scala/com/microsoft/tpch/TpchFunctionalQueries.scala b/benchmark/scala/src/main/scala/com/microsoft/tpch/TpchFunctionalQueries.scala
@@ -250,10 +250,10 @@ class TpchFunctionalQueries(spark: SparkSession, tpchRoot: String)
 
     customer.join(order, $"c_custkey" === order("o_custkey")
       && !special(order("o_comment")), "left_outer")
-      .groupBy($"o_custkey")
+      .groupBy($"c_custkey")
       .agg(count($"o_orderkey").as("c_count"))
       .groupBy($"c_count")
-      .agg(count($"o_custkey").as("custdist"))
+      .agg(count($"*").as("custdist"))
       .sort($"custdist".desc, $"c_count".desc)
       .show()
   }