You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@carbondata.apache.org by "Xinyu Zeng (Jira)" <ji...@apache.org> on 2022/05/09 05:29:00 UTC
[jira] [Updated] (CARBONDATA-4334) Blocklet skipping not works?
[ https://issues.apache.org/jira/browse/CARBONDATA-4334?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Xinyu Zeng updated CARBONDATA-4334:
-----------------------------------
Description:
Hi, I follow the instructions of <Installing and Configuring CarbonData to run locally with Spark Shell> in Quick Start guide, using apache-carbondata-2.3.0-bin-spark2.3.4-hadoop2.7.2.jar.
The scala script I am running is :
{code:java}
import java.io.File
import org.apache.spark.sql.{CarbonEnv, SaveMode, SparkSession}
import org.apache.carbondata.core.constants.{CarbonCommonConstants, CarbonV3DataFormatConstants}
import org.apache.carbondata.core.util.CarbonProperties
def createSparkSession(): SparkSession =
{
CarbonProperties.getInstance() .addProperty(CarbonCommonConstants.ENABLE_UNSAFE_COLUMN_PAGE, "true") .addProperty(CarbonV3DataFormatConstants.BLOCKLET_SIZE_IN_MB, "64") .addProperty(CarbonCommonConstants.CARBON_BADRECORDS_LOC, "") valspark=SparkSession.builder().config(sc.getConf).enableHiveSupport.config("spark.sql.extensions","org.apache.spark.sql.CarbonExtensions").getOrCreate() CarbonEnv.getInstance(spark) spark.sparkContext.setLogLevel("ERROR") spark }
val spark = createSparkSession()
spark.sql("DROP TABLE IF EXISTS lineitem_0509")
spark.sql(s""" | CREATE TABLE IF NOT EXISTS lineitem_0509( | L_ORDERKEY BIGINT, | L_PARTKEY BIGINT, | L_SUPPKEY BIGINT, | L_LINENUMBER INTEGER, | L_QUANTITY DECIMAL, | L_EXTENDEDPRICE DECIMAL, | L_DISCOUNT DECIMAL, | L_TAX DECIMAL, | L_RETURNFLAG CHAR(1), | L_LINESTATUS CHAR(1), | L_SHIPDATE DATE, | L_COMMITDATE DATE, | L_RECEIPTDATE DATE, | L_SHIPINSTRUCT CHAR(25), | L_SHIPMODE CHAR(10), | L_COMMENT VARCHAR(44)) | STORED AS carbondata |TBLPROPERTIES ('SORT_COLUMNS'='L_SHIPDATE', 'COLUMN_META_CACHE'='L_SHIPDATE') """.stripMargin)spark.sql("LOAD DATA INPATH '/root/lineitem_1M_shipdate.csv' INTO TABLE lineitem_0509")
spark.sql("LOAD DATA INPATH '/root/lineitem_1M_shipdate.csv' INTO TABLE lineitem_0509")
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
// this is used to implicitly convert an RDD to a DataFrame.
import sqlContext.implicits._
spark.time
{ val df = sqlContext.sql("SELECT l_extendedprice, l_discount, l_quantity FROM lineitem_0509 WHERE l_shipdate >= to_date('1994-01-01') AND l_shipdate < to_date('1995-01-01')") } {code}
However, it seems like the blocklet index is not effective because whether I add the shipdate filter or not the select query takes the same time to run
was:
Hi, I follow the instructions of <Installing and Configuring CarbonData to run locally with Spark Shell> in Quick Start guide, using apache-carbondata-2.3.0-bin-spark2.3.4-hadoop2.7.2.jar.
The scala script I am running is :
```scala
import java.io.File
import org.apache.spark.sql.\{CarbonEnv, SaveMode, SparkSession}
import org.apache.carbondata.core.constants.\{CarbonCommonConstants, CarbonV3DataFormatConstants}
import org.apache.carbondata.core.util.CarbonProperties
def createSparkSession(): SparkSession = {
CarbonProperties.getInstance()
.addProperty(CarbonCommonConstants.ENABLE_UNSAFE_COLUMN_PAGE, "true")
.addProperty(CarbonV3DataFormatConstants.BLOCKLET_SIZE_IN_MB, "64")
.addProperty(CarbonCommonConstants.CARBON_BADRECORDS_LOC, "")
valspark=SparkSession.builder().config(sc.getConf).enableHiveSupport.config("spark.sql.extensions","org.apache.spark.sql.CarbonExtensions").getOrCreate()
CarbonEnv.getInstance(spark)
spark.sparkContext.setLogLevel("ERROR")
spark
}
val spark = createSparkSession()
spark.sql(s"CREATE TABLE LINEITEM_STD (L_ORDERKEY BIGINT,L_PARTKEY BIGINT,L_SUPPKEYBIGINT, L_LINENUMBER INTEGER,L_QUANTITY double,L_EXTENDEDPRICEdouble,L_DISCOUNT double,L_TAX double,L_RETURNFLAG string,L_LINESTATUSstring,L_SHIPDATE DATE,L_COMMITDATE DATE,L_RECEIPTDATEDATE,L_SHIPINSTRUCT string,L_SHIPMODE string,L_COMMENT string) STORED AS 'carbondata' TBLPROPERTIES ('SORT_COLUMNS'='','table_blocklet_size'='90','table_blocksize'='120')")
// spark.sql("DROP TABLE IF EXISTS lineitem_0509")
// spark.sql(s"""
// | CREATE TABLE IF NOT EXISTS lineitem_0509(
// | L_ORDERKEY BIGINT,
// | L_PARTKEY BIGINT,
// | L_SUPPKEY BIGINT,
// | L_LINENUMBER INTEGER,
// | L_QUANTITY DECIMAL,
// | L_EXTENDEDPRICE DECIMAL,
// | L_DISCOUNT DECIMAL,
// | L_TAX DECIMAL,
// | L_RETURNFLAG CHAR(1),
// | L_LINESTATUS CHAR(1),
// | L_SHIPDATE DATE,
// | L_COMMITDATE DATE,
// | L_RECEIPTDATE DATE,
// | L_SHIPINSTRUCT CHAR(25),
// | L_SHIPMODE CHAR(10),
// | L_COMMENT VARCHAR(44))
// | STORED AS carbondata
// |TBLPROPERTIES ('SORT_COLUMNS'='L_SHIPDATE', 'COLUMN_META_CACHE'='L_SHIPDATE')
// """.stripMargin)
spark.sql("LOAD DATA INPATH '/root/lineitem_1M_shipdate.csv' INTO TABLE lineitem_0509")
val sqlContext = new org.apache.spark.sql.SQLContext(sc)
// this is used to implicitly convert an RDD to a DataFrame.
import sqlContext.implicits._
spark.time{
val df = sqlContext.sql("SELECT l_extendedprice, l_discount, l_quantity FROM lineitem_0509 WHERE l_shipdate >= to_date('1994-01-01') AND l_shipdate < to_date('1995-01-01')")
}
```
However, it seems like the blocklet index is not effective because whether I add the shipdate filter or not the select query takes the same time to run
> Blocklet skipping not works?
> ----------------------------
>
> Key: CARBONDATA-4334
> URL: https://issues.apache.org/jira/browse/CARBONDATA-4334
> Project: CarbonData
> Issue Type: Improvement
> Reporter: Xinyu Zeng
> Priority: Major
>
> Hi, I follow the instructions of <Installing and Configuring CarbonData to run locally with Spark Shell> in Quick Start guide, using apache-carbondata-2.3.0-bin-spark2.3.4-hadoop2.7.2.jar.
> The scala script I am running is :
>
> {code:java}
> import java.io.File
> import org.apache.spark.sql.{CarbonEnv, SaveMode, SparkSession}
> import org.apache.carbondata.core.constants.{CarbonCommonConstants, CarbonV3DataFormatConstants}
> import org.apache.carbondata.core.util.CarbonProperties
> def createSparkSession(): SparkSession =
> {
> CarbonProperties.getInstance() .addProperty(CarbonCommonConstants.ENABLE_UNSAFE_COLUMN_PAGE, "true") .addProperty(CarbonV3DataFormatConstants.BLOCKLET_SIZE_IN_MB, "64") .addProperty(CarbonCommonConstants.CARBON_BADRECORDS_LOC, "") valspark=SparkSession.builder().config(sc.getConf).enableHiveSupport.config("spark.sql.extensions","org.apache.spark.sql.CarbonExtensions").getOrCreate() CarbonEnv.getInstance(spark) spark.sparkContext.setLogLevel("ERROR") spark }
> val spark = createSparkSession()
> spark.sql("DROP TABLE IF EXISTS lineitem_0509")
> spark.sql(s""" | CREATE TABLE IF NOT EXISTS lineitem_0509( | L_ORDERKEY BIGINT, | L_PARTKEY BIGINT, | L_SUPPKEY BIGINT, | L_LINENUMBER INTEGER, | L_QUANTITY DECIMAL, | L_EXTENDEDPRICE DECIMAL, | L_DISCOUNT DECIMAL, | L_TAX DECIMAL, | L_RETURNFLAG CHAR(1), | L_LINESTATUS CHAR(1), | L_SHIPDATE DATE, | L_COMMITDATE DATE, | L_RECEIPTDATE DATE, | L_SHIPINSTRUCT CHAR(25), | L_SHIPMODE CHAR(10), | L_COMMENT VARCHAR(44)) | STORED AS carbondata |TBLPROPERTIES ('SORT_COLUMNS'='L_SHIPDATE', 'COLUMN_META_CACHE'='L_SHIPDATE') """.stripMargin)spark.sql("LOAD DATA INPATH '/root/lineitem_1M_shipdate.csv' INTO TABLE lineitem_0509")
> spark.sql("LOAD DATA INPATH '/root/lineitem_1M_shipdate.csv' INTO TABLE lineitem_0509")
> val sqlContext = new org.apache.spark.sql.SQLContext(sc)
> // this is used to implicitly convert an RDD to a DataFrame.
> import sqlContext.implicits._
> spark.time
> { val df = sqlContext.sql("SELECT l_extendedprice, l_discount, l_quantity FROM lineitem_0509 WHERE l_shipdate >= to_date('1994-01-01') AND l_shipdate < to_date('1995-01-01')") } {code}
>
> However, it seems like the blocklet index is not effective because whether I add the shipdate filter or not the select query takes the same time to run
--
This message was sent by Atlassian Jira
(v8.20.7#820007)