Spark TPCDS Data Gen

开启 Spark-Shell

$SPARK_HOME/bin/spark-shell --master local[10] --jars {PATH}/spark-sql-perf-1.2/target/scala-2.12/spark-sql-perf_2.12-0.5.1-SNAPSHOT.jar

Gen Data

Gen TCPDS Parquet

val tools_path = "/opt/Beaver/tpcds-kit/tools"
val data_path = "hdfs://{IP}:9000/tpcds_parquet_tpcds_kit_1_0/1"
val database_name = "tpcds_parquet_tpcds_kit_1_0_scale_1_db"
val scale = "1"
val p = scale.toInt / 2048.0
val catalog_returns_p = (263 * p + 1).toInt
val catalog_sales_p = (2285 * p * 0.5 * 0.5 + 1).toInt
val store_returns_p = (429 * p + 1).toInt
val store_sales_p = (3164 * p * 0.5 * 0.5 + 1).toInt
val web_returns_p = (198 * p + 1).toInt
val web_sales_p = (1207 * p * 0.5 * 0.5 + 1).toInt
val format = "parquet"
val codec = "snappy"

val useDoubleForDecimal = false

val partitionTables = false
val clusterByPartitionColumns = partitionTables

import com.databricks.spark.sql.perf.tpcds.TPCDSTables
spark.sqlContext.setConf(s"spark.sql.$format.compression.codec", codec)

val tables = new TPCDSTables(spark, spark.sqlContext, tools_path, scale, useDoubleForDecimal)

tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "call_center", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "catalog_page", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "customer", 6)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "customer_address", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "customer_demographics", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "date_dim", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "household_demographics", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "income_band", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "inventory", 6)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "item", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "promotion", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "reason", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "ship_mode", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "store", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "time_dim", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "warehouse", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "web_page", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "web_site", 1)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "catalog_sales", catalog_sales_p)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "catalog_returns", catalog_returns_p)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "store_sales", store_sales_p)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "store_returns", store_returns_p)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "web_sales", web_sales_p)
tables.genData(data_path, format, true, partitionTables, clusterByPartitionColumns, false, "web_returns", web_returns_p)

tables.createExternalTables(data_path, format, database_name, overwrite = true, discoverPartitions = partitionTables)

Gen TPCH ORC

import com.databricks.spark.sql.perf.tpch._

val tools_path = "/opt/Beaver/tpch-dbgen"
val format = "orc"
val useDoubleForDecimal = false
val partitionTables = false
val scaleFactor = "1"

val data_path = s"hdfs://{IP}:9000/tpch_${format}_${scaleFactor}"

val numPartitions =1
val databaseName = s"tpch_${format}_${scaleFactor}_db"


val clusterByPartitionColumns = partitionTables

val tables = new TPCHTables(spark, spark.sqlContext,
    dbgenDir = tools_path,
    scaleFactor = scaleFactor,
    useDoubleForDecimal = useDoubleForDecimal,
    useStringForDate = false)

spark.sqlContext.setConf("spark.sql.files.maxRecordsPerFile", "200000000")

tables.genData(
    location = data_path,
    format = format,
    overwrite = true, // overwrite the data that is already there
    partitionTables, // do not create the partitioned fact tables
    clusterByPartitionColumns, // shuffle to get partitions coalesced into single files.
    filterOutNullPartitionValues = false, // true to filter out the partition with NULL key value
    tableFilter = "", // "" means generate all tables
    numPartitions = numPartitions) // how many dsdgen partitions to run - number of input tasks.

// Create the specified database
sql(s"drop database if exists $databaseName CASCADE")
sql(s"create database $databaseName")
// 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(data_path, format, databaseName, overwrite = true, discoverPartitions = false)

创建Metadata

Parquet create database/tables

val tools_path = "/opt/Beaver/tpcds-kit/tools"
val data_path = "hdfs://10.1.2.206:9000/user/sparkuser/part_tpcds_decimal_1000/"
val database_name = "sr242_parquet_part_tpcds_decimal_1000"
val scale = "1000"
val useDoubleForDecimal = false
val format = "parquet"
val partitionTables = true
import com.databricks.spark.sql.perf.tpcds.TPCDSTables

val tables = new TPCDSTables(spark, spark.sqlContext, tools_path, scale, useDoubleForDecimal)
tables.createExternalTables(data_path, format, database_name, overwrite = true, discoverPartitions = partitionTables)

Arrow create database/tables

val data_path= "hdfs://{IP}:9000/{PATH}/part_tpcds_decimal_1000/"
val databaseName = "arrow_part_tpcds_decimal_1000"
val tables = Seq("call_center", "catalog_page", "catalog_returns", "catalog_sales", "customer", "customer_address", "customer_demographics", "date_dim", "household_demographics", "income_band", "inventory", "item", "promotion", "reason", "ship_mode", "store", "store_returns", "store_sales", "time_dim", "warehouse", "web_page", "web_returns", "web_sales", "web_site")
val partitionTables = true
spark.sql(s"DROP database if exists $databaseName CASCADE")
if (spark.catalog.databaseExists(s"$databaseName")) {
	println(s"$databaseName has exists!")
}else{
	spark.sql(s"create database if not exists $databaseName").show
	spark.sql(s"use $databaseName").show
	for (table <- tables) {
		if (spark.catalog.tableExists(s"$table")){
			println(s"$table has exists!")
		}else{
			spark.catalog.createTable(s"$table", s"$data_path/$table", "arrow")
		}
	}
	if (partitionTables) {
            for (table <- tables) {
                try{
                    spark.sql(s"ALTER TABLE $table RECOVER PARTITIONS").show
                }catch{
                        case e: Exception => println(e)
                }
            }
        }
}

使用ALTER 修改meta 信息

val data_path= "hdfs://{IP}:9000/{PATH}/part_tpcds_decimal_1000/"
val databaseName = "parquet_part_tpcds_decimal_1000"
val tables = Seq("call_center", "catalog_page", "catalog_returns", "catalog_sales", "customer", "customer_address", "customer_demographics", "date_dim", "household_demographics", "income_band", "inventory", "item", "promotion", "reason", "ship_mode", "store", "store_returns", "store_sales", "time_dim", "warehouse", "web_page", "web_returns", "web_sales", "web_site")

spark.sql(s"use $databaseName").show
for (table <- tables) {
    try{
        spark.sql(s"ALTER TABLE $table SET LOCATION '$data_path/$table'").show
    }catch{
            case e: Exception => println(e)
    }
}