Merge branch 'master' into kubernetes_state

Author: turboFei

docs/client/python/index.rst

@@ -21,5 +21,4 @@ Python
     :maxdepth: 2
 
     pyhive
-    pyspark
     jaydebeapi

docs/client/python/pyspark.md

@@ -15,27 +15,131 @@
 - limitations under the License.
 -->
 
-# Hive Dialect Support
+# Hive JDBC Data Source Dialect
 
-Hive Dialect plugin aims to provide Hive Dialect support to Spark's JDBC source.
+Hive JDBC Data Source dialect plugin aims to provide Hive Dialect support to [Spark's JDBC Data Source](https://spark.apache.org/docs/latest/sql-data-sources-jdbc.html).
 It will auto registered to Spark and applied to JDBC sources with url prefix of `jdbc:hive2://` or `jdbc:kyuubi://`.
 
-Hive Dialect helps to solve failures access Kyuubi. It fails and unexpected results when querying data from Kyuubi as JDBC source with Hive JDBC Driver or Kyuubi Hive JDBC Driver in Spark, as Spark JDBC provides no Hive Dialect support out of box and quoting columns and other identifiers in ANSI as "table.column" rather than in HiveSQL style as \`table\`.\`column\`.
+Hive Dialect helps to solve failures access Kyuubi. It fails and unexpected results when querying data from Kyuubi as
+JDBC data source with Hive JDBC Driver or Kyuubi Hive JDBC Driver in Spark, as Spark JDBC provides no Hive Dialect support
+out of box and quoting columns and other identifiers in ANSI as "table.column" rather than in HiveSQL style as
+\`table\`.\`column\`.
+
+Notes: this is an inefficient way to access data stored in Hive warehouse, you can see more discussions at [SPARK-47482](https://github.com/apache/spark/pull/45609).
 
 ## Features
 
-- quote identifier in Hive SQL style
+- Quote identifier in Hive SQL style
+
+  e.g. Quote `table.column` in \`table\`.\`column\`
+
+- Adapt to Hive data type definitions
+
+  Reference: https://cwiki.apache.org/confluence/display/hive/languagemanual+types
+
+## Preparation
+
+### Prepare JDBC driver
+
+Prepare JDBC driver jar file. Supported Hive compatible JDBC Driver as below:
+
+|                          Driver                           |            Driver Class Name            |                                                 Remarks                                                  |
+|-----------------------------------------------------------|-----------------------------------------|----------------------------------------------------------------------------------------------------------|
+| Kyuubi Hive JDBC Driver ([doc](../jdbc/kyuubi_jdbc.html)) | org.apache.kyuubi.jdbc.KyuubiHiveDriver | Use v1.6.1 or later versions, which includes [KYUUBI #3484](https://github.com/apache/kyuubi/pull/3485). |
+| Hive JDBC Driver ([doc](../jdbc/hive_jdbc.html))          | org.apache.hive.jdbc.HiveDriver         | The Hive JDBC driver is already included in official Spark binary distribution.                          |
+
+Refer to docs of the driver and prepare the JDBC driver jar file.
+
+### Prepare JDBC Hive Dialect extension
+
+Prepare the plugin jar file `kyuubi-extension-spark-jdbc-dialect_-*.jar`.
+
+Get the Kyuubi Hive Dialect Extension jar from Maven Central
+
+```
+<dependency>
+    <groupId>org.apache.kyuubi</groupId>
+    <artifactId>kyuubi-extension-spark-jdbc-dialect_2.12</artifactId>
+    <version>{latest-version}</version>
+</dependency>
+```
+
+Or, compile the extension by executing
+
+```
+build/mvn clean package -pl :kyuubi-extension-spark-jdbc-dialect_2.12 -DskipTests
+```
 
-  eg. Quote `table.column` in \`table\`.\`column\`
+then get the extension jar under `extensions/spark/kyuubi-extension-spark-jdbc-dialect/target`.
+
+If you like, you can compile the extension jar with the corresponding Maven's profile on you compile command,
+i.e. you can get extension jar for Spark 3.5 by compiling with `-Pspark-3.5`
+
+### Including jars of JDBC driver and Hive Dialect extension
+
+Choose one of the following ways to include jar files in Spark.
+
+- Put the jar file of JDBC driver and Hive Dialect to `$SPARK_HOME/jars` directory to make it visible for all Spark applications. And adding `spark.sql.extensions = org.apache.spark.sql.dialect.KyuubiSparkJdbcDialectExtension` to `$SPARK_HOME/conf/spark_defaults.conf.`
+
+- With each `spark-submit`(or `spark-sql`, `pyspark` etc.) commands, include the JDBC driver when submitting the application with `--packages`, and the Hive Dialect plugins with `--jars`
+
+```
+$SPARK_HOME/bin/spark-submit \
+  --packages org.apache.hive:hive-jdbc:x.y.z \
+  --jars /path/kyuubi-extension-spark-jdbc-dialect_-*.jar \
+  ...
+```
+
+- Setting jars and config with SparkSession builder
+
+```
+val spark = SparkSession.builder
+    .config("spark.jars", "/path/hive-jdbc-x.y.z.jar,/path/kyuubi-extension-spark-jdbc-dialect_-*.jar")
+    .config("spark.sql.extensions", "org.apache.spark.sql.dialect.KyuubiSparkJdbcDialectExtension")
+    .getOrCreate()
+```
 
 ## Usage
 
-1. Get the Kyuubi Hive Dialect Extension jar
-   1. compile the extension by executing `build/mvn clean package -pl :kyuubi-extension-spark-jdbc-dialect_2.12 -DskipTests`
-   2. get the extension jar under `extensions/spark/kyuubi-extension-spark-jdbc-dialect/target`
-   3. If you like, you can compile the extension jar with the corresponding Maven's profile on you compile command, i.e. you can get extension jar for Spark 3.5 by compiling with `-Pspark-3.5`
-2. Put the Kyuubi Hive Dialect Extension jar `kyuubi-extension-spark-jdbc-dialect_-*.jar` into `$SPARK_HOME/jars`
-3. Enable `KyuubiSparkJdbcDialectExtension`, by setting `spark.sql.extensions=org.apache.spark.sql.dialect.KyuubiSparkJdbcDialectExtension`, i.e.
-   - add a config into `$SPARK_HOME/conf/spark-defaults.conf`
-   - or add setting config in SparkSession builder
+### Using as JDBC Datasource programmingly
+
+```
+# Loading data from Kyuubi via HiveDriver as JDBC datasource
+val jdbcDF = spark.read
+    .format("jdbc")
+    .option("driver", "org.apache.hive.jdbc.HiveDriver")
+    .option("url", "jdbc:hive2://kyuubi_server_ip:port")
+    .option("dbtable", "schema.tablename")
+    .option("user", "username")
+    .option("password", "password")
+    .option("query", "select * from testdb.src_table")
+    .load()
+```
+
+### Using as JDBC Datasource table with SQL
+
+From Spark 3.2.0, [`CREATE DATASOURCE TABLE`](https://spark.apache.org/docs/latest/sql-ref-syntax-ddl-create-table-datasource.html)
+is supported to create jdbc source with SQL.
+
+```sql
+-- create JDBC data source table
+CREATE TABLE kyuubi_table
+USING JDBC
+OPTIONS (
+    driver='org.apache.hive.jdbc.HiveDriver',
+    url='jdbc:hive2://kyuubi_server_ip:port',
+    user='user',
+    password='password',
+    dbtable='testdb.some_table'
+)
+
+-- query data
+SELECT * FROM kyuubi_table
+
+-- write data in overwrite mode
+INSERT OVERWRITE kyuubi_table SELECT ...
+
+-- write data in append mode
+INSERT INTO kyuubi_table SELECT ...
+```
 

docs/extensions/engines/spark/jdbc-dialect.md

@@ -67,6 +67,9 @@ public int findColumn(String columnName) throws SQLException {
   @Override
   public BigDecimal getBigDecimal(int columnIndex) throws SQLException {
     final Object val = getObject(columnIndex);
+    if (val == null) {
+      return null;
+    }
     if (val instanceof BigDecimal) {
       return (BigDecimal) val;
     }

kyuubi-hive-jdbc/src/main/java/org/apache/kyuubi/jdbc/hive/KyuubiBaseResultSet.java

@@ -26,13 +26,14 @@ import com.fasterxml.jackson.module.scala.DefaultScalaModule
 import org.apache.kyuubi.Logging
 import org.apache.kyuubi.config.KyuubiConf
 import org.apache.kyuubi.engine.ApplicationState.ApplicationState
+import org.apache.kyuubi.server.metadata.MetadataManager
 
 trait ApplicationOperation {
 
   /**
    * Step for initializing the instance.
    */
-  def initialize(conf: KyuubiConf): Unit
+  def initialize(conf: KyuubiConf, metadataManager: Option[MetadataManager]): Unit
 
   /**
    * Step to clean up the instance

kyuubi-server/src/main/scala/org/apache/kyuubi/engine/ApplicationOperation.scala

@@ -22,13 +22,14 @@ import java.nio.file.Paths
 import scala.sys.process._
 
 import org.apache.kyuubi.config.KyuubiConf
+import org.apache.kyuubi.server.metadata.MetadataManager
 
 class JpsApplicationOperation extends ApplicationOperation {
   import ApplicationOperation._
 
   private var runner: String = _
 
-  override def initialize(conf: KyuubiConf): Unit = {
+  override def initialize(conf: KyuubiConf, metadataManager: Option[MetadataManager]): Unit = {
     val jps = sys.env.get("JAVA_HOME").orElse(sys.props.get("java.home"))
       .map(Paths.get(_, "bin", "jps").toString)
       .getOrElse("jps")

kyuubi-server/src/main/scala/org/apache/kyuubi/engine/JpsApplicationOperation.scala

@@ -37,9 +37,8 @@ import org.apache.kyuubi.config.KyuubiConf.KubernetesCleanupDriverPodStrategy.{A
 import org.apache.kyuubi.engine.ApplicationState.{isTerminated, ApplicationState, FAILED, FINISHED, KILLED, NOT_FOUND, PENDING, RUNNING, UNKNOWN}
 import org.apache.kyuubi.engine.KubernetesResourceEventTypes.KubernetesResourceEventType
 import org.apache.kyuubi.operation.OperationState
-import org.apache.kyuubi.server.KyuubiServer
+import org.apache.kyuubi.server.metadata.MetadataManager
 import org.apache.kyuubi.server.metadata.api.KubernetesEngineInfo
-import org.apache.kyuubi.session.KyuubiSessionManager
 import org.apache.kyuubi.util.{KubernetesUtils, ThreadUtils}
 
 class KubernetesApplicationOperation extends ApplicationOperation with Logging {
@@ -81,8 +80,7 @@ class KubernetesApplicationOperation extends ApplicationOperation with Logging {
     kubernetesClients.computeIfAbsent(kubernetesInfo, kInfo => buildKubernetesClient(kInfo))
   }
 
-  private def metadataManager = KyuubiServer.kyuubiServer.backendService
-    .sessionManager.asInstanceOf[KyuubiSessionManager].metadataManager
+  private var metadataManager: Option[MetadataManager] = _
 
   // Visible for testing
   private[engine] def checkKubernetesInfo(kubernetesInfo: KubernetesInfo): Unit = {
@@ -120,8 +118,9 @@ class KubernetesApplicationOperation extends ApplicationOperation with Logging {
     }
   }
 
-  override def initialize(conf: KyuubiConf): Unit = {
+  override def initialize(conf: KyuubiConf, metadataManager: Option[MetadataManager]): Unit = {
     kyuubiConf = conf
+    this.metadataManager = metadataManager
     info("Start initializing Kubernetes application operation.")
     submitTimeout = conf.get(KyuubiConf.ENGINE_KUBERNETES_SUBMIT_TIMEOUT)
     // Defer cleaning terminated application information

kyuubi-server/src/main/scala/org/apache/kyuubi/engine/KubernetesApplicationOperation.scala

@@ -31,10 +31,12 @@ import org.apache.kyuubi.config.KyuubiConf
 import org.apache.kyuubi.engine.KubernetesApplicationOperation.LABEL_KYUUBI_UNIQUE_KEY
 import org.apache.kyuubi.engine.flink.FlinkProcessBuilder
 import org.apache.kyuubi.engine.spark.SparkProcessBuilder
+import org.apache.kyuubi.server.metadata.MetadataManager
 import org.apache.kyuubi.service.AbstractService
 import org.apache.kyuubi.util.reflect.ReflectUtils._
 
-class KyuubiApplicationManager extends AbstractService("KyuubiApplicationManager") {
+class KyuubiApplicationManager(metadataManager: Option[MetadataManager])
+  extends AbstractService("KyuubiApplicationManager") {
 
   // TODO: maybe add a configuration is better
   private val operations =
@@ -43,7 +45,7 @@ class KyuubiApplicationManager extends AbstractService("KyuubiApplicationManager
   override def initialize(conf: KyuubiConf): Unit = {
     operations.foreach { op =>
       try {
-        op.initialize(conf)
+        op.initialize(conf, metadataManager)
       } catch {
         case NonFatal(e) => warn(s"Error starting ${op.getClass.getSimpleName}: ${e.getMessage}")
       }

kyuubi-server/src/main/scala/org/apache/kyuubi/engine/KyuubiApplicationManager.scala

@@ -32,6 +32,7 @@ import org.apache.kyuubi.config.KyuubiConf.YarnUserStrategy._
 import org.apache.kyuubi.engine.ApplicationOperation._
 import org.apache.kyuubi.engine.ApplicationState.ApplicationState
 import org.apache.kyuubi.engine.YarnApplicationOperation.toApplicationState
+import org.apache.kyuubi.server.metadata.MetadataManager
 import org.apache.kyuubi.util.KyuubiHadoopUtils
 
 class YarnApplicationOperation extends ApplicationOperation with Logging {
@@ -40,7 +41,7 @@ class YarnApplicationOperation extends ApplicationOperation with Logging {
   @volatile private var adminYarnClient: Option[YarnClient] = None
   private var submitTimeout: Long = _
 
-  override def initialize(conf: KyuubiConf): Unit = {
+  override def initialize(conf: KyuubiConf, metadataManager: Option[MetadataManager]): Unit = {
     submitTimeout = conf.get(KyuubiConf.ENGINE_YARN_SUBMIT_TIMEOUT)
     yarnConf = KyuubiHadoopUtils.newYarnConfiguration(conf)