This tutorial provides a quick introduction to use CarbonData. To follow along with this guide, download a packaged release of CarbonData from the CarbonData website. Alternatively, it can be created following Building CarbonData steps.
CarbonData supports Spark versions up to 2.4. Please download Spark package from Spark website
Create a sample.csv file using the following commands. The CSV file is required for loading data into CarbonData
cd carbondata
cat > sample.csv << EOF
id,name,city,age
1,david,shenzhen,31
2,eason,shenzhen,27
3,jarry,wuhan,35
EOF
CarbonData can be integrated with Spark, Presto, Flink and Hive execution engines. The below documentation guides on Installing and Configuring with these execution engines.
Installing and Configuring CarbonData to run locally with Spark SQL CLI
Installing and Configuring CarbonData to run locally with Spark Shell
Installing and Configuring CarbonData on Standalone Spark Cluster
Installing and Configuring CarbonData on Spark on YARN Cluster
Installing and Configuring CarbonData Thrift Server for Query Execution
Installing and Configuring CarbonData on Presto
Installing and Configuring CarbonData on Hive
CarbonData supports read and write with HDFS
CarbonData supports read and write with S3
CarbonData supports read and write with Alluxio
This will work with spark 2.3+ versions. In Spark SQL CLI, it uses CarbonExtensions to customize the SparkSession with CarbonData's parser, analyzer, optimizer and physical planning strategy rules in Spark. To enable CarbonExtensions, we need to add the following configuration.
| Key | Value |
|---|---|
| spark.sql.extensions | org.apache.spark.sql.CarbonExtensions |
Start Spark SQL CLI by running the following command in the Spark directory:
./bin/spark-sql --conf spark.sql.extensions=org.apache.spark.sql.CarbonExtensions --jars <carbondata assembly jar path>
CREATE TABLE IF NOT EXISTS test_table (
id string,
name string,
city string,
age Int)
STORED AS carbondata;
NOTE: CarbonExtensions only support "STORED AS carbondata" and "USING carbondata"
LOAD DATA INPATH '/local-path/sample.csv' INTO TABLE test_table;
LOAD DATA INPATH 'hdfs://hdfs-path/sample.csv' INTO TABLE test_table;
insert into table test_table select '1', 'name1', 'city1', 1;
NOTE: Please provide the real file path of sample.csv for the above script.
If you get "tablestatus.lock" issue, please refer to FAQ
SELECT * FROM test_table;
SELECT city, avg(age), sum(age)
FROM test_table
GROUP BY city;
Apache Spark Shell provides a simple way to learn the API, as well as a powerful tool to analyze data interactively. Please visit Apache Spark Documentation for more details on the Spark shell.
Start Spark shell by running the following command in the Spark directory:
./bin/spark-shell --jars <carbondata assembly jar path>
NOTE: Path where packaged release of CarbonData was downloaded or assembly jar will be available after building CarbonData and can be copied from ./assembly/target/scala-2.1x/apache-carbondata_xxx.jar
In this shell, SparkSession is readily available as spark and Spark context is readily available as sc.
In order to create a CarbonSession we will have to configure it explicitly in the following manner :
import org.apache.spark.sql.SparkSession
import org.apache.spark.sql.CarbonSession._
val carbon = SparkSession.builder().config(sc.getConf).getOrCreateCarbonSession("<carbon_store_path>")
NOTE
../carbon.metastore, user can provide own metastore location to CarbonSession like
SparkSession.builder().config(sc.getConf).getOrCreateCarbonSession("<carbon_store_path>", "<local metastore path>").<carbon_store_path>, like /carbon/data/store, hdfs://localhost:9000/carbon/data/store or s3a://carbon/data/store.Start Spark shell by running the following command in the Spark directory:
./bin/spark-shell --conf spark.sql.extensions=org.apache.spark.sql.CarbonExtensions --jars <carbondata assembly jar path>
NOTE
spark instead of carbon.
We also can create a new SparkSession instead of the built-in SparkSession spark if need.
It need to add "org.apache.spark.sql.CarbonExtensions" into spark configuration "spark.sql.extensions".
SparkSession newSpark = SparkSession
.builder()
.config(sc.getConf)
.enableHiveSupport
.config("spark.sql.extensions","org.apache.spark.sql.CarbonExtensions")
.getOrCreate()
carbon.sql(
s"""
| CREATE TABLE IF NOT EXISTS test_table(
| id string,
| name string,
| city string,
| age Int)
| STORED AS carbondata
""".stripMargin)
NOTE: The following table list all supported syntax:
| create table | SparkSession with CarbonExtensions | CarbonSession |
|---|---|---|
| STORED AS carbondata | yes | yes |
| USING carbondata | yes | yes |
| STORED BY 'carbondata' | no | yes |
| STORED BY 'org.apache.carbondata.format' | no | yes |
We suggest to use CarbonExtensions instead of CarbonSession.
carbon.sql("LOAD DATA INPATH '/path/to/sample.csv' INTO TABLE test_table")
NOTE: Please provide the real file path of sample.csv for the above script.
If you get "tablestatus.lock" issue, please refer to FAQ
carbon.sql("SELECT * FROM test_table").show()
carbon.sql(
s"""
| SELECT city, avg(age), sum(age)
| FROM test_table
| GROUP BY city
""".stripMargin).show()
Build the CarbonData project and get the assembly jar from ./assembly/target/scala-2.1x/apache-carbondata_xxx.jar.
Copy ./assembly/target/scala-2.1x/apache-carbondata_xxx.jar to $SPARK_HOME/carbonlib folder.
NOTE: Create the carbonlib folder if it does not exist inside $SPARK_HOME path.
Add the carbonlib folder path in the Spark classpath. (Edit $SPARK_HOME/conf/spark-env.sh file and modify the value of SPARK_CLASSPATH by appending $SPARK_HOME/carbonlib/* to the existing value)
Copy the ./conf/carbon.properties.template file from CarbonData repository to $SPARK_HOME/conf/ folder and rename the file to carbon.properties.
Repeat Step 2 to Step 5 in all the nodes of the cluster.
In Spark node[master], configure the properties mentioned in the following table in $SPARK_HOME/conf/spark-defaults.conf file.
| Property | Value | Description |
|---|---|---|
| spark.driver.extraJavaOptions | -Dcarbon.properties.filepath = $SPARK_HOME/conf/carbon.properties |
A string of extra JVM options to pass to the driver. For instance, GC settings or other logging. |
| spark.executor.extraJavaOptions | -Dcarbon.properties.filepath = $SPARK_HOME/conf/carbon.properties |
A string of extra JVM options to pass to executors. For instance, GC settings or other logging. NOTE: You can enter multiple values separated by space. |
./bin/spark-shell \
--master spark://HOSTNAME:PORT \
--total-executor-cores 2 \
--executor-memory 2G
NOTE:
This section provides the procedure to install CarbonData on "Spark on YARN" cluster.
The following steps are only for Driver Nodes. (Driver nodes are the one which starts the spark context.)
Build the CarbonData project and get the assembly jar from ./assembly/target/scala-2.1x/apache-carbondata_xxx.jar and copy to $SPARK_HOME/carbonlib folder.
NOTE: Create the carbonlib folder if it does not exists inside $SPARK_HOME path.
Copy the ./conf/carbon.properties.template file from CarbonData repository to $SPARK_HOME/conf/ folder and rename the file to carbon.properties.
Create tar.gz file of carbonlib folder and move it inside the carbonlib folder.
cd $SPARK_HOME
tar -zcvf carbondata.tar.gz carbonlib/
mv carbondata.tar.gz carbonlib/
$SPARK_HOME/conf/spark-defaults.conf file.| Property | Description | Value |
|---|---|---|
| spark.master | Set this value to run the Spark in yarn cluster mode. | Set yarn-client to run the Spark in yarn cluster mode. |
| spark.yarn.dist.files | Comma-separated list of files to be placed in the working directory of each executor. | $SPARK_HOME/conf/carbon.properties |
| spark.yarn.dist.archives | Comma-separated list of archives to be extracted into the working directory of each executor. | $SPARK_HOME/carbonlib/carbondata.tar.gz |
| spark.executor.extraJavaOptions | A string of extra JVM options to pass to executors. For instance NOTE: You can enter multiple values separated by space. | -Dcarbon.properties.filepath = carbon.properties |
| spark.executor.extraClassPath | Extra classpath entries to prepend to the classpath of executors. NOTE: If SPARK_CLASSPATH is defined in spark-env.sh, then comment it and append the values in below parameter spark.driver.extraClassPath | carbondata.tar.gz/carbonlib/* |
| spark.driver.extraClassPath | Extra classpath entries to prepend to the classpath of the driver. NOTE: If SPARK_CLASSPATH is defined in spark-env.sh, then comment it and append the value in below parameter spark.driver.extraClassPath. | $SPARK_HOME/carbonlib/* |
| spark.driver.extraJavaOptions | A string of extra JVM options to pass to the driver. For instance, GC settings or other logging. | -Dcarbon.properties.filepath = $SPARK_HOME/conf/carbon.properties |
./bin/spark-shell \
--master yarn-client \
--driver-memory 1G \
--executor-memory 2G \
--executor-cores 2
NOTE:
a. cd $SPARK_HOME
b. Run the following command to start the CarbonData thrift server.
./bin/spark-submit \
--class org.apache.carbondata.spark.thriftserver.CarbonThriftServer \
$SPARK_HOME/carbonlib/$CARBON_ASSEMBLY_JAR
| Parameter | Description | Example |
|---|---|---|
| CARBON_ASSEMBLY_JAR | CarbonData assembly jar name present in the $SPARK_HOME/carbonlib/ folder. |
apache-carbondata-xx.jar |
c. Run the following command to work with S3 storage.
./bin/spark-submit \
--class org.apache.carbondata.spark.thriftserver.CarbonThriftServer \
$SPARK_HOME/carbonlib/$CARBON_ASSEMBLY_JAR <access_key> <secret_key> <endpoint>
| Parameter | Description | Example |
|---|---|---|
| CARBON_ASSEMBLY_JAR | CarbonData assembly jar name present in the $SPARK_HOME/carbonlib/ folder. |
apache-carbondata-xx.jar |
| access_key | Access key for S3 storage | |
| secret_key | Secret key for S3 storage | |
| endpoint | Endpoint for connecting to S3 storage |
NOTE: From Spark 1.6, by default the Thrift server runs in multi-session mode. Which means each JDBC/ODBC connection owns a copy of their own SQL configuration and temporary function registry. Cached tables are still shared though. If you prefer to run the Thrift server in single-session mode and share all SQL configuration and temporary function registry, please set option spark.sql.hive.thriftServer.singleSession to true. You may either add this option to spark-defaults.conf, or pass it to spark-submit.sh via --conf:
./bin/spark-submit \
--conf spark.sql.hive.thriftServer.singleSession=true \
--class org.apache.carbondata.spark.thriftserver.CarbonThriftServer \
$SPARK_HOME/carbonlib/$CARBON_ASSEMBLY_JAR
But in single-session mode, if one user changes the database from one connection, the database of the other connections will be changed too.
Examples
./bin/spark-submit \
--class org.apache.carbondata.spark.thriftserver.CarbonThriftServer \
$SPARK_HOME/carbonlib/apache-carbondata-xxx.jar
./bin/spark-submit \
--class org.apache.carbondata.spark.thriftserver.CarbonThriftServer \
--num-executors 3 \
--driver-memory 20G \
--executor-memory 250G \
--executor-cores 32 \
$SPARK_HOME/carbonlib/apache-carbondata-xxx.jar
cd $SPARK_HOME
./sbin/start-thriftserver.sh
./bin/beeline -u jdbc:hive2://<thriftserver_host>:port
Example
./bin/beeline -u jdbc:hive2://10.10.10.10:10000
NOTE: CarbonData tables cannot be created nor loaded from Presto. User needs to create CarbonData Table and load data into it either with Spark or SDK or C++ SDK. Once the table is created, it can be queried from Presto.
Please refer the presto guide linked below.
prestodb guide - prestodb
prestosql guide - prestosql
Once installed the presto with carbonData as per the above guide, you can use the Presto CLI on the coordinator to query data sources in the catalog using the Presto workers.
List the schemas(databases) available
show schemas;
Selected the schema where CarbonData table resides
use carbonschema;
List the available tables
show tables;
Query from the available tables
select * from carbon_table;
Note: Create Tables and data loads should be done before executing queries as we can not create carbon table from this interface.