Export - CSV (All fields)
Export - CSV (Current fields)
SPARK-49898
Correct documentation and default for event log task metrics accumulator logging flag from SPARK-42204
Don't use deprecated Hive JDBC client parameters in docs / testsSPARK-49647
Pick up system property defaults in SharedSparkContext test suite mixin
In CountVectorizer, only read binary parameter once per transform, not once per row
Improve performance of ParserUtils.unescapeSQLString
Improve performance of SparkConf.get(ConfigEntry)
Add opt-in flag to have SparkSubmit automatically call System.exit after user code main method exits
Use static regex Pattern instances in common/utils JavaUtils
Fix nullability of try_parse_jsonSPARK-48128
BitwiseCount / bit_count generated code for boolean inputs fails to compile
Fix ClassCastException in NTile.checkInputDataTypes() when argument is non-foldable or of wrong type
Fix flaky pyspark.sql.dataframe.DataFrame.writeStream doctest by stopping streaming query
Avoid noisy RejectedExecutionExceptions during StandaloneSchedulerBackend shutdown
Add "--add-opens=java.base/jdk.internal.ref=ALL-UNNAMED" so Platform can access cleaner on Java 9+
Reduce memory usage of LiveStageMetrics.accumIdsToMetricType
Use await() instead of awaitUninterruptibly() in TransportClientFactory.createClient()
Fix flakiness in Kafka RDD suites due to port binding configuration issue
Spark 3.4 history server's SQL tab incorrectly groups SQL executions when replaying event logs from Spark 3.3 and earlier
Omit "Task Executor Metrics" field in JsonProtocol output if values are all zero
Remove logging of Accumulables in Task/Stage start events in JsonProtocol
Remove redundant logging of TaskMetrics internal accumulators in JsonProtocol event logs
JsonProtocol should skip logging of push-based shuffle read metrics when push-based shuffle is disabled
Fix wrong child call in SQLShuffleWriteMetricsReporter.decRecordsWritten()
Use interruptible lock instead of synchronized in TransportClientFactory.createClient()
Use interruptible lock instead of synchronized in Executor.updateDependencies()
Avoid noisy warnings logs when spark.scheduler.listenerbus.metrics.maxListenerClassesTimed = 0
Reconsider design of ignoreCorruptFiles feature
ExecutionListenerManager's registration of the ExecutionListenerBus should be lazy
Race condition related to interruption of task threads while they are in RocksDBLoader.loadLibrary()
If spark.default.parallelism is unset, RDD defaultPartitioner may pick a value that is too large to successfully run
Reduce memory overhead of TransportCipher$EncryptedMessage's byteRawChannel buffer
Reconsider exposure of Json4s symbols in public ResourceInformation.toJson API
Fix multiple small bugs in JsonProtocol, impacting StorageLevel and Task/Executor resource requests
Improve EventLoggingListener and ReplayListener performance by replacing Json4S ASTs with Jackson trees
SHOW CREATE TABLE should suggest 'AS SERDE' for Hive tables with unsupported serde configurations
Stop using Log4J2's extended throwable logging pattern in default logging configurations
Cleanup unused spark.rpc.numRetries and spark.rpc.retry.wait configs
CodeGenerator.addBufferedState() does not properly handle UDTs
Multiple optimizations related to CurrentOrigin.withOrigin
Cache LogicalPlan.isStreaming() in a lazy val
Add tree pattern pruning to CTESubstitution rule
Add checks to DebugFilesystem to ensure that FS operations are not performed in the DAGScheduler event loop
Spark 3.2.0 POMs are no longer "dependency reduced"
Use SparkSubmitTestUtils to core and use it in SparkSubmitSuite
Fix incorrect use of assume() in tests
Use latest URL for Google Cloud Storage mirror of Maven Central
Optimizations for arrays and maps in ParquetRowConverter
Avoid unnecessary WrappedArray roundtrip in GenericArrayData constructor
Avoid unnecessary InternalRow copies in ParquetRowConverter
"TaskResultLost (result lost from block manager)" error message is misleading in case result fetch is caused by client-side network connectivity issues