Fix the seed to 100 and use distribute by instead of order by.

README.md

@@ -16,7 +16,7 @@ import com.databricks.spark.sql.perf.tpcds.Tables
 // dsdgenDir is the location of dsdgen tool installed in your machines.
 val tables = new Tables(sqlContext, dsdgenDir, scaleFactor)
 // Generate data.
-tables.genData(location, format, overwrite, partitionTables, useDoubleForDecimal, orderByPartitionColumns)
+tables.genData(location, format, overwrite, partitionTables, useDoubleForDecimal, clusterByPartitionColumns)
 // Create metastore tables in a specified database for your data.
 // Once tables are created, the current database will be switched to the specified database.
 tables.createExternalTables(location, format, databaseName, overwrite)

src/main/scala/com/databricks/spark/sql/perf/tpcds/Tables.scala

@@ -51,10 +51,11 @@ class Tables(sqlContext: SQLContext, dsdgenDir: String, scaleFactor: Int) extend
             sys.error(s"Could not find dsdgen at $dsdgen or /$dsdgen. Run install")
           }
 
+          // Note: RNGSEED is the RNG seed used by the data generator. Right now, it is fixed to 100.
           val parallel = if (partitions > 1) s"-parallel $partitions -child $i" else ""
           val commands = Seq(
             "bash", "-c",
-            s"cd $localToolsDir && ./dsdgen -table $name -filter Y -scale $scaleFactor $parallel")
+            s"cd $localToolsDir && ./dsdgen -table $name -filter Y -scale $scaleFactor -RNGSEED 100 $parallel")
           println(commands)
           commands.lines
         }
@@ -103,18 +104,39 @@ class Tables(sqlContext: SQLContext, dsdgenDir: String, scaleFactor: Int) extend
       Table(name, partitionColumns, newFields:_*)
     }
 
-    def genData(location: String, format: String, overwrite: Boolean, orderByPartitionColumns: Boolean): Unit = {
+    def genData(location: String, format: String, overwrite: Boolean, clusterByPartitionColumns: Boolean): Unit = {
       val mode = if (overwrite) SaveMode.Overwrite else SaveMode.Ignore
 
+      val data = df
+      val tempTableName = s"${name}_text"
+      data.registerTempTable(tempTableName)
+
       val writer = if (partitionColumns.nonEmpty) {
-        if (orderByPartitionColumns) {
-          df.orderBy(partitionColumns.map(col): _*).write
+        if (clusterByPartitionColumns) {
+          val columnString = data.schema.fields.map { field =>
+            field.name
+          }.mkString(",")
+          val partitionColumnString = partitionColumns.mkString(",")
+
+          val query =
+            s"""
+              |SELECT
+              |  $columnString
+              |FROM
+              |  $tempTableName
+              |DISTRIBUTED BY
+              |  $partitionColumnString
+            """.stripMargin
+          val grouped = sqlContext.sql(query)
+          println(s"Pre-clustering with partitioning columns with query $query.")
+          logInfo(s"Pre-clustering with partitioning columns with query $query.")
+          grouped.write
         } else {
-          df.write
+          data.write
         }
       } else {
         // If the table is not partitioned, coalesce the data to a single file.
-        df.coalesce(1).write
+        data.coalesce(1).write
       }
       writer.format(format).mode(mode)
       if (partitionColumns.nonEmpty) {
@@ -123,6 +145,7 @@ class Tables(sqlContext: SQLContext, dsdgenDir: String, scaleFactor: Int) extend
       println(s"Generating table $name in database to $location with save mode $mode.")
       logInfo(s"Generating table $name in database to $location with save mode $mode.")
       writer.save(location)
+      sqlContext.dropTempTable(tempTableName)
     }
 
     def createExternalTable(location: String, format: String, databaseName: String, overwrite: Boolean): Unit = {
@@ -151,7 +174,7 @@ class Tables(sqlContext: SQLContext, dsdgenDir: String, scaleFactor: Int) extend
       overwrite: Boolean,
       partitionTables: Boolean,
       useDoubleForDecimal: Boolean,
-      orderByPartitionColumns: Boolean): Unit = {
+      clusterByPartitionColumns: Boolean): Unit = {
     val tablesToBeGenerated = if (partitionTables) {
       tables
     } else {
@@ -166,7 +189,7 @@ class Tables(sqlContext: SQLContext, dsdgenDir: String, scaleFactor: Int) extend
 
     withSpecifiedDataType.foreach { table =>
       val tableLocation = s"$location/${table.name}"
-      table.genData(tableLocation, format, overwrite, orderByPartitionColumns)
+      table.genData(tableLocation, format, overwrite, clusterByPartitionColumns)
     }
   }