[SPARK-19669][SQL] Open up visibility for sharedState, sessionState, and a few other functions

## What changes were proposed in this pull request?
To ease debugging, most of Spark SQL internals have public level visibility. Two of the most important internal states, sharedState and sessionState, however, are package private. It would make more sense to open these up as well with clear documentation that they are internal.

In addition, users currently have way to set active/default SparkSession, but no way to actually get them back. We should open those up as well.

## How was this patch tested?
N/A - only visibility change.

Author: Reynold Xin <rxin@databricks.com>

Closes #17002 from rxin/SPARK-19669.
This commit is contained in:
Reynold Xin 2017-02-20 12:21:07 -08:00 committed by gatorsmile
parent ead4ba0eb5
commit 0733a54a45
3 changed files with 27 additions and 8 deletions

View file

@ -28,7 +28,7 @@ import org.apache.spark.util.Utils
* logging messages at different levels using methods that only evaluate parameters lazily if the
* log level is enabled.
*/
private[spark] trait Logging {
trait Logging {
// Make the log field transient so that objects with Logging can
// be serialized and used on another machine

View file

@ -95,18 +95,28 @@ class SparkSession private(
/**
* State shared across sessions, including the `SparkContext`, cached data, listener,
* and a catalog that interacts with external systems.
*
* This is internal to Spark and there is no guarantee on interface stability.
*
* @since 2.2.0
*/
@InterfaceStability.Unstable
@transient
private[sql] lazy val sharedState: SharedState = {
lazy val sharedState: SharedState = {
existingSharedState.getOrElse(new SharedState(sparkContext))
}
/**
* State isolated across sessions, including SQL configurations, temporary tables, registered
* functions, and everything else that accepts a [[org.apache.spark.sql.internal.SQLConf]].
*
* This is internal to Spark and there is no guarantee on interface stability.
*
* @since 2.2.0
*/
@InterfaceStability.Unstable
@transient
private[sql] lazy val sessionState: SessionState = {
lazy val sessionState: SessionState = {
SparkSession.reflect[SessionState, SparkSession](
SparkSession.sessionStateClassName(sparkContext.conf),
self)
@ -613,7 +623,6 @@ class SparkSession private(
*
* @since 2.1.0
*/
@InterfaceStability.Stable
def time[T](f: => T): T = {
val start = System.nanoTime()
val ret = f
@ -928,9 +937,19 @@ object SparkSession {
defaultSession.set(null)
}
private[sql] def getActiveSession: Option[SparkSession] = Option(activeThreadSession.get)
/**
* Returns the active SparkSession for the current thread, returned by the builder.
*
* @since 2.2.0
*/
def getActiveSession: Option[SparkSession] = Option(activeThreadSession.get)
private[sql] def getDefaultSession: Option[SparkSession] = Option(defaultSession.get)
/**
* Returns the default SparkSession that is returned by the builder.
*
* @since 2.2.0
*/
def getDefaultSession: Option[SparkSession] = Option(defaultSession.get)
/** A global SQL listener used for the SQL UI. */
private[sql] val sqlListener = new AtomicReference[SQLListener]()

View file

@ -39,7 +39,7 @@ private[sql] class SharedState(val sparkContext: SparkContext) extends Logging {
// Load hive-site.xml into hadoopConf and determine the warehouse path we want to use, based on
// the config from both hive and Spark SQL. Finally set the warehouse config value to sparkConf.
val warehousePath = {
val warehousePath: String = {
val configFile = Utils.getContextOrSparkClassLoader.getResource("hive-site.xml")
if (configFile != null) {
sparkContext.hadoopConfiguration.addResource(configFile)
@ -103,7 +103,7 @@ private[sql] class SharedState(val sparkContext: SparkContext) extends Logging {
/**
* A manager for global temporary views.
*/
val globalTempViewManager = {
val globalTempViewManager: GlobalTempViewManager = {
// System preserved database should not exists in metastore. However it's hard to guarantee it
// for every session, because case-sensitivity differs. Here we always lowercase it to make our
// life easier.