Databricks Runtime 14.3 LTS

The following release notes provide information about Databricks Runtime 14.3 LTS, powered by Apache Spark 3.5.0.

Databricks released this version in February 2024.

New features and improvements

• Workspace file system support in PySpark User-Defined Functions (UDFs) on shared clusters
• Support for deletion vector MERGE optimizations without Photon
• Spark catalog APIs now fully supported in shared access mode
• Delta UniForm is now generally available
• New SQL function EXECUTE IMMEDIATE
• Recompute data skipping statistics for Delta tables
• Query state information for stateful streaming queries
• Use Microsoft Entra ID for Kafka authentication on shared clusters
• Support added for file and partition pruning to improve query performance
• Declare temporary variables in a SQL Session
• Thriftserver updates to remove unused features
• Use truststore and keystore files in Unity Catalog volumes
• Native XML file format support (Public Preview)
• Support for Cloudflare R2 storage (Public Preview)
• Spark and dbutils access to workspace files support on shared-access Unity Catalog clusters
• Init scripts and cluster library support on shared-access Unity Catalog clusters

Workspace file system support in PySpark User-Defined Functions (UDFs) on shared clusters

PySpark UDFs on shared clusters can now import Python modules from Git folders, workspace files, or UC volumes.

For more information about working with modules in Git folders or workspace files, see Work with Python and R modules.

Support for deletion vector MERGE optimizations without Photon

Photon is no longer required for MERGE operations to leverage deletion vectors optimizations. See What are deletion vectors?.

Spark catalog APIs now fully supported in shared access mode

You can now use all functions in the spark.catalog API in both Python and Scala on compute configured with shared access mode.

Delta UniForm is now generally available

UniForm is now generally available and uses the IcebergCompatV2 table feature. You can now enable or upgrade UniForm on existing tables. See Use UniForm to read Delta tables with Iceberg clients.

New SQL function EXECUTE IMMEDIATE

You can now use the EXECUTE IMMEDIATE syntax to support parameterized queries in SQL. See EXECUTE IMMEDIATE.

Recompute data skipping statistics for Delta tables

You can now recompute statistics stored in the Delta log after changing columns used for data skipping. See Specify Delta statistics columns.

Query state information for stateful streaming queries

You can now query Structured Streaming state data and metadata. See Read Structured Streaming state information.

Use Microsoft Entra ID for Kafka authentication on shared clusters

You can now authenticate Event Hubs services via OAuth with Microsoft Entra ID on compute configured with shared access mode. See Service Principal authentication with Microsoft Entra ID and Azure Event Hubs.

Support added for file and partition pruning to improve query performance

To speed up some queries that rely on null-tolerant equality in JOIN conditions, we now support DynamicFilePruning and DynamicPartitionPruning for the EqualNullSafe operator in JOINs.

Declare temporary variables in a SQL Session

This release introduces the ability to declare temporary variables in a session which can be set and then referred to from within queries. See Variables.

Thriftserver updates to remove unused features

The Thriftserver code has been updated to remove code for obsolete features. Because of these changes, the following configurations are no longer supported:

• The Hive auxiliary JARs, configured using the hive.aux.jars.path property, are not supported anymore for hive-thriftserver connections.
• The Hive global init file (.hiverc), whose location is configured using the hive.server2.global.init.file.location property or the HIVE_CONF_DIR environment variable, is not supported anymore for hive-thriftserver connections.

Use truststore and keystore files in Unity Catalog volumes

You can now use truststore and keystore files in Unity Catalog volumes to authenticate to a Confluent Schema Registry for avro or protocol buffer data. See documentation for avro or protocol buffers.

Native XML file format support (Public Preview)

Native XML file format support is now in Public Preview. XML file format support enables ingestion, querying, and parsing of XML data for batch processing or streaming. It can automatically infer and evolve schema and data types, supports SQL expressions like from_xml, and can generate XML documents. It doesn’t require external jars and works seamlessly with Auto Loader, read_files, COPY INTO, and Delta Live Tables. See Read and write XML files.

Support for Cloudflare R2 storage (Public Preview)

You can now use Cloudflare R2 as cloud storage for data registered in Unity Catalog. Cloudflare R2 is intended primarily for Delta Sharing use cases in which you want to avoid the data egress fees charged by cloud providers when data crosses regions. R2 storage supports all of the Databricks data and AI assets supported in AWS S3, Azure Data Lake Storage Gen2, and Google Cloud Storage. See Use Cloudflare R2 replicas or migrate storage to R2 and Create a storage credential for connecting to Cloudflare R2.

Spark and dbutils access to workspace files support on shared-access Unity Catalog clusters

Spark and dbutils read and write access to workspace files is now supported on Unity Catalog clusters in shared access mode. See Work with workspace files.

Init scripts and cluster library support on shared-access Unity Catalog clusters

Installation of cluster-scoped init scripts and Python and JAR libraries on Unity Catalog clusters in shared access mode, including installation using cluster policies, is now generally available. Databricks recommends installing init scripts and libraries from Unity Catalog volumes.

Library upgrades

• Upgraded Python libraries:
  • fastjsonschema from 2.19.0 to 2.19.1
  • filelock from 3.12.4 to 3.13.1
  • googleapis-common-protos from 1.61.0 to 1.62.0
  • packaging from 22.0 to 23.2
• Upgraded R libraries:
  • foreign from 0.8-82 to 0.8-85
  • nlme from 3.1-162 to 3.1-163
  • rpart from 4.1.19 to 4.1.21
• Upgraded Java libraries:
  • com.databricks.databricks-sdk-java from 0.7.0 to 0.13.0
  • org.apache.orc.orc-core from 1.9.1-shaded-protobuf to 1.9.2-shaded-protobuf
  • org.apache.orc.orc-mapreduce from 1.9.1-shaded-protobuf to 1.9.2-shaded-protobuf
  • org.apache.orc.orc-shims from 1.9.1 to 1.9.2
  • org.scala-lang.modules.scala-collection-compat_2.12 from 2.9.0 to 2.11.0

Apache Spark

Databricks Runtime 14.3 includes Apache Spark 3.5.0. This release includes all Spark fixes and improvements included in Databricks Runtime 14.2, as well as the following additional bug fixes and improvements made to Spark:

• [SPARK-46541] [SC-153546][SQL][CONNECT] Fix the ambiguous column reference in self join
• [SPARK-45433] Revert “[SC-145163][SQL] Fix CSV/JSON schema infer…
• [SPARK-46723] [14.3][SASP-2792][SC-153425][CONNECT][SCALA] Make addArtifact retryable
• [SPARK-46660] [SC-153391][CONNECT] ReattachExecute requests updates aliveness of SessionHolder
• [SPARK-46670] [SC-153273][PYTHON][SQL] Make DataSourceManager self clone-able by separating static and runtime Python Data Sources
• [SPARK-46720] [SC-153410][SQL][PYTHON] Refactor Python Data Source to align with other DSv2 built-in Data Sources
• [SPARK-46684] [SC-153275][PYTHON][CONNECT] Fix CoGroup.applyInPandas/Arrow to pass arguments properly
• [SPARK-46667] [SC-153271][SC-153263][SQL] XML: Throw error on multiple XML data source
• [SPARK-46382] [SC-151881][SQL]XML: Default ignoreSurroundingSpaces to true
• [SPARK-46382] [SC-153178][SQL] XML: Update doc for ignoreSurroundingSpaces
• [SPARK-45292] Revert “[SC-151609][SQL][HIVE] Remove Guava from shared classes from IsolatedClientLoader”
• [SPARK-45292] [SC-151609][SQL][HIVE] Remove Guava from shared classes from IsolatedClientLoader
• [SPARK-46311] [SC-150137][CORE] Log the final state of drivers during Master.removeDriver
• [SPARK-46413] [SC-151052][PYTHON] Validate returnType of Arrow Python UDF
• [SPARK-46633] [WARMFIX][SC-153092][SQL] Fix Avro reader to handle zero-length blocks
• [SPARK-46537] [SC-151286][SQL] Convert NPE and asserts from commands to internal errors
• [SPARK-46179] [SC-151678][SQL] Add CrossDbmsQueryTestSuites, which runs other DBMS against golden files with other DBMS, starting with Postgres
• [SPARK-44001] [SC-151413][PROTOBUF] Add option to allow unwrapping protobuf well known wrapper types
• [SPARK-40876] [SC-151786][SQL] Widening type promotion for decimals with larger scale in Parquet readers
• [SPARK-46605] [SC-151769][CONNECT] Make lit/typedLit function in connect module support s.c.immutable.ArraySeq
• [SPARK-46634] [SC-153005][SQL] literal validation should not drill down to null fields
• [SPARK-37039] [SC-153094][PS] Fix Series.astype to work properly with missing value
• [SPARK-46312] [SC-150163][CORE] Use lower_camel_case in store_types.proto
• [SPARK-46630] [SC-153084][SQL] XML: Validate XML element name on write
• [SPARK-46627] [SC-152981][SS][UI] Fix timeline tooltip content on streaming ui
• [SPARK-46248] [SC-151774][SQL] XML: Support for ignoreCorruptFiles and ignoreMissingFiles options
• [SPARK-46386] [SC-150766][PYTHON] Improve assertions of observation (pyspark.sql.observation)
• [SPARK-46581] [SC-151789][CORE] Update comment on isZero in AccumulatorV2
• [SPARK-46601] [SC-151785] [CORE] Fix log error in handleStatusMessage
• [SPARK-46568] [SC-151685][PYTHON] Make Python data source options a case-insensitive dictionary
• [SPARK-46611] [SC-151783][CORE] Remove ThreadLocal by replace SimpleDateFormat with DateTimeFormatter
• [SPARK-46604] [SC-151768][SQL] Make Literal.apply support s.c.immuable.ArraySeq
• [SPARK-46324] [SC-150223][SQL][PYTHON] Fix the output name of pyspark.sql.functions.user and session_user
• [SPARK-46621] [SC-151794][PYTHON] Address null from Exception.getMessage in Py4J captured exception
• [SPARK-46598] [SC-151767][SQL] OrcColumnarBatchReader should respect the memory mode when creating column vectors for the missing column
• [SPARK-46613] [SC-151778][SQL][PYTHON] Log full exception when failed to lookup Python Data Sources
• [SPARK-46559] [SC-151364][MLLIB] Wrap the export in the package name with backticks
• [SPARK-46522] [SC-151784][PYTHON] Block Python data source registration with name conflicts
• [SPARK-45580] [SC-149956][SQL] Handle case where a nested subquery becomes an existence join
• [SPARK-46609] [SC-151714][SQL] Avoid exponential explosion in PartitioningPreservingUnaryExecNode
• [SPARK-46535] [SC-151288][SQL] Fix NPE when describe extended a column without col stats
• [SPARK-46599] [SC-147661][SC-151770][SQL] XML: Use TypeCoercion.findTightestCommonType for compatibility check
• [SPARK-40876] [SC-151129][SQL] Widening type promotions in Parquet readers
• [SPARK-46179] [SC-151069][SQL] Pull out code into reusable functions in SQLQueryTestSuite
• [SPARK-46586] [SC-151679][SQL] Support s.c.immutable.ArraySeq as customCollectionCls in MapObjects
• [SPARK-46585] [SC-151622][CORE] Directly constructed metricPeaks as an immutable.ArraySeq instead of use mutable.ArraySeq.toSeq in Executor
• [SPARK-46488] [SC-151173][SQL] Skipping trimAll call during timestamp parsing
• [SPARK-46231] [SC-149724][PYTHON] Migrate all remaining NotImplementedError & TypeError into PySpark error framework
• [SPARK-46348] [SC-150281][CORE] Support spark.deploy.recoveryTimeout
• [SPARK-46313] [SC-150164][CORE] Log Spark HA recovery duration
• [SPARK-46358] [SC-150307][CONNECT] Simplify the condition check in the ResponseValidator#verifyResponse
• [SPARK-46380] [SC-151078][SQL]Replace current time/date prior to evaluating inline table expressions.
• [SPARK-46563] [SC-151441][SQL] View simpleString not follow conf spark.sql.debug.maxToStringFields
• [SPARK-46101] [SC-149211][CORE][SQL][MLLIB][SS][R][CONNCT][GRAPHX] Reduce stack depth by replace (string|array).size with (string|array).length
• [SPARK-46539] [SC-151469][SQL] SELECT * EXCEPT(all fields from a struct) results in an assertion failure
• [SPARK-46565] [SC-151414][PYTHON] Refine error classes and error messages for Python data sources
• [SPARK-46576] [SC-151614][SQL] Improve error messages for unsupported data source save mode
• [SPARK-46540] [SC-151355][PYTHON] Respect column names when Python data source read function outputs named Row objects
• [SPARK-46577] [SC-151448][SQL] HiveMetastoreLazyInitializationSuite leaks hive’s SessionState
• [SPARK-44556] [SC-151562][SQL] Reuse OrcTail when enable vectorizedReader
• [SPARK-46587] [SC-151618][SQL] XML: Fix XSD big integer conversion
• [SPARK-46382] [SC-151297][SQL] XML: Capture values interspersed between elements
• [SPARK-46567] [SC-151447][CORE] Remove ThreadLocal for ReadAheadInputStream
• [SPARK-45917] [SC-151269][PYTHON][SQL] Automatic registration of Python Data Source on startup
• [SPARK-28386] [SC-151014][SQL] Cannot resolve ORDER BY columns with GROUP BY and HAVING
• [SPARK-46524] [SC-151446][SQL] Improve error messages for invalid save mode
• [SPARK-46294] [SC-150672][SQL] Clean up semantics of init vs zero value
• [SPARK-46301] [SC-150100][CORE] Support spark.worker.(initial|max)RegistrationRetries
• [SPARK-46542] [SC-151324][SQL] Remove the check for c>=0 from ExternalCatalogUtils#needsEscaping as it is always true
• [SPARK-46553] [SC-151360][PS] FutureWarning for interpolate with object dtype
• [SPARK-45914] [SC-151312][PYTHON] Support commit and abort API for Python data source write
• [SPARK-46543] [SC-151350][PYTHON][CONNECT] Make json_tuple throw PySparkValueError for empty fields
• [SPARK-46520] [SC-151227][PYTHON] Support overwrite mode for Python data source write
• [SPARK-46502] [SC-151235][SQL] Support timestamp types in UnwrapCastInBinaryComparison
• [SPARK-46532] [SC-151279][CONNECT] Pass message parameters in metadata of ErrorInfo
• [SPARK-46397] Revert “[SC-151311][PYTHON][CONNECT] Function sha2 should raise PySparkValueError for invalid numBits”
• [SPARK-46170] [SC-149472][SQL] Support inject adaptive query post planner strategy rules in SparkSessionExtensions
• [SPARK-46444] [SC-151195][SQL] V2SessionCatalog#createTable should not load the table
• [SPARK-46397] [SC-151311][PYTHON][CONNECT] Function sha2 should raise PySparkValueError for invalid numBits
• [SPARK-46145] [SC-149471][SQL] spark.catalog.listTables does not throw exception when the table or view is not found
• [SPARK-46272] [SC-151012][SQL] Support CTAS using DSv2 sources
• [SPARK-46480] [SC-151127][CORE][SQL] Fix NPE when table cache task attempt
• [SPARK-46100] [SC-149051][CORE][PYTHON] Reduce stack depth by replace (string|array).size with (string|array).length
• [SPARK-45795] [SC-150764][SQL] DS V2 supports push down Mode
• [SPARK-46485] [SC-151133][SQL] V1Write should not add Sort when not needed
• [SPARK-46468] [SC-151121] [SQL] Handle COUNT bug for EXISTS subqueries with Aggregate without grouping keys
• [SPARK-46246] [SC-150927][SQL] EXECUTE IMMEDIATE SQL support
• [SPARK-46498] [SC-151199][CORE] Remove shuffleServiceEnabled from o.a.spark.util.Utils#getConfiguredLocalDirs
• [SPARK-46440] [SC-150807][SQL] Set the rebase configs to the CORRECTED mode by default
• [SPARK-45525] [SC-151120][SQL][PYTHON] Support for Python data source write using DSv2
• [SPARK-46505] [SC-151187][CONNECT] Make bytes threshold configurable in ProtoUtils.abbreviate
• [SPARK-46447] [SC-151025][SQL] Remove the legacy datetime rebasing SQL configs
• [SPARK-46443] [SC-151123][SQL] Decimal precision and scale should decided by H2 dialect.
• [SPARK-46384] [SC-150694][SPARK-46404][SS][UI] Fix Operation Duration Stack Chart on Structured Streaming Page
• [SPARK-46207] [SC-151068][SQL] Support MergeInto in DataFrameWriterV2
• [SPARK-46452] [SC-151018][SQL] Add a new API in DataWriter to write an iterator of records
• [SPARK-46273] [SC-150313][SQL] Support INSERT INTO/OVERWRITE using DSv2 sources
• [SPARK-46344] [SC-150267][CORE] Warn properly when a driver exists successfully but master is disconnected
• [SPARK-46284] [SC-149944][PYTHON][CONNECT] Add session_user function to Python
• [SPARK-46043] [SC-149786][SQL] Support create table using DSv2 sources
• [SPARK-46323] [SC-150191][PYTHON] Fix the output name of pyspark.sql.functions.now
• [SPARK-46465] [SC-151059][PYTHON][CONNECT] Add Column.isNaN in PySpark
• [SPARK-46456] [SC-151050][CORE] Add spark.ui.jettyStopTimeout to set Jetty server stop timeout to unblock SparkContext shutdown
• [SPARK-43427] [SC-150356][PROTOBUF] spark protobuf: allow upcasting unsigned integer types
• [SPARK-46399] [SC-151013][14.x][Core] Add exit status to the Application End event for the use of Spark Listener
• [SPARK-46423] [SC-150752][PYTHON][SQL] Make the Python Data Source instance at DataSource.lookupDataSourceV2
• [SPARK-46424] [SC-150765][PYTHON][SQL] Support Python metrics in Python Data Source
• [SPARK-46330] [SC-151015] Loading of Spark UI blocks for a long time when HybridStore enabled
• [SPARK-46378] [SC-150397][SQL] Still remove Sort after converting Aggregate to Project
• [SPARK-45506] [SC-146959][CONNECT] Add ivy URI support to SparkConnect addArtifact
• [SPARK-45814] [SC-147871][CONNECT][SQL]Make ArrowConverters.createEmptyArrowBatch call close() to avoid memory leak
• [SPARK-46427] [SC-150759][PYTHON][SQL] Change Python Data Source’s description to be pretty in explain
• [SPARK-45597] [SC-150730][PYTHON][SQL] Support creating table using a Python data source in SQL (DSv2 exec)
• [SPARK-46402] [SC-150700][PYTHON] Add getMessageParameters and getQueryContext support
• [SPARK-46453] [SC-150897][CONNECT] Throw exception from internalError() in SessionHolder
• [SPARK-45758] [SC-147454][SQL] Introduce a mapper for hadoop compression codecs
• [SPARK-46213] [PYTHON] Introduce PySparkImportError for error framework
• [SPARK-46230] [SC-149960][PYTHON] Migrate RetriesExceeded into PySpark error
• [SPARK-45035] [SC-145887][SQL] Fix ignoreCorruptFiles/ignoreMissingFiles with multiline CSV/JSON will report error
• [SPARK-46289] [SC-150846][SQL] Support ordering UDTs in interpreted mode
• [SPARK-46229] [SC-150798][PYTHON][CONNECT] Add applyInArrow to groupBy and cogroup in Spark Connect
• [SPARK-46420] [SC-150847][SQL] Remove unused transport in SparkSQLCLIDriver
• [SPARK-46226] [PYTHON] Migrate all remaining RuntimeError into PySpark error framework
• [SPARK-45796] [SC-150613][SQL] Support MODE() WITHIN GROUP (ORDER BY col)
• [SPARK-40559] [SC-149686][PYTHON][14.X] Add applyInArrow to groupBy and cogroup
• [SPARK-46069] [SC-149672][SQL] Support unwrap timestamp type to date type
• [SPARK-46406] [SC-150770][SQL] Assign a name to the error class _LEGACY_ERROR_TEMP_1023
• [SPARK-46431] [SC-150782][SS] Convert IllegalStateException to internalError in session iterators
• [SPARK-45807] [SC-150751][SQL] Improve ViewCatalog API
• [SPARK-46009] [SC-149771][SQL][CONNECT] Merge the parse rule of PercentileCont and PercentileDisc into functionCall
• [SPARK-46403] [SC-150756][SQL] Decode parquet binary with getBytesUnsafe method
• [SPARK-46389] [SC-150779][CORE] Manually close the RocksDB/LevelDB instance when checkVersion throw Exception
• [SPARK-46360] [SC-150376][PYTHON] Enhance error message debugging with new getMessage API
• [SPARK-46233] [SC-149766][PYTHON] Migrate all remaining AttributeError into PySpark error framework
• [SPARK-46394] [SC-150599][SQL] Fix spark.catalog.listDatabases() issues on schemas with special characters when spark.sql.legacy.keepCommandOutputSchema set to true
• [SPARK-45964] [SC-148484][SQL] Remove private sql accessor in XML and JSON package under catalyst package
• [SPARK-45856] [SC-148691] Move ArtifactManager from Spark Connect into SparkSession (sql/core)
• [SPARK-46409] [SC-150714][CONNECT] Fix spark-connect-scala-client launch script
• [SPARK-46416] [SC-150699][CORE] Add @tailrec to HadoopFSUtils#shouldFilterOutPath
• [SPARK-46115] [SC-149154][SQL] Restrict charsets in encode()
• [SPARK-46253] [SC-150416][PYTHON] Plan Python data source read using MapInArrow
• [SPARK-46339] [SC-150266][SS] Directory with batch number name should not be treated as metadata log
• [SPARK-46353] [SC-150396][CORE] Refactor to improve RegisterWorker unit test coverage
• [SPARK-45826] [SC-149061][SQL] Add a SQL config for stack traces in DataFrame query context
• [SPARK-45649] [SC-150300][SQL] Unify the prepare framework for OffsetWindowFunctionFrame
• [SPARK-46346] [SC-150282][CORE] Fix Master to update a worker from UNKNOWN to ALIVE on RegisterWorker msg
• [SPARK-46388] [SC-150496][SQL] HiveAnalysis misses the pattern guard of query.resolved
• [SPARK-46401] [SC-150609][CORE] Use !isEmpty() on RoaringBitmap instead of getCardinality() > 0 in RemoteBlockPushResolver
• [SPARK-46393] [SC-150603][SQL] Classify exceptions in the JDBC table catalog
• [SPARK-45798] [SC-150513][CONNECT] Followup: add serverSessionId to SessionHolderInfo
• [SPARK-46153] [SC-146564][SC-150508][SQL] XML: Add TimestampNTZType support
• [SPARK-46075] [SC-150393][CONNECT] Improvements to SparkConnectSessionManager
• [SPARK-46357] [SC-150596] Replace incorrect documentation use of setConf with conf.set
• [SPARK-46202] [SC-150420][CONNECT] Expose new ArtifactManager APIs to support custom target directories
• [SPARK-45824] [SC-147648][SQL] Enforce the error class in ParseException
• [SPARK-45845] [SC-148163][SS][UI] Add number of evicted state rows to streaming UI
• [SPARK-46316] [SC-150181][CORE] Enable buf-lint-action on core module
• [SPARK-45816] [SC-147656][SQL] Return NULL when overflowing during casting from timestamp to integers
• [SPARK-46333] [SC-150263][SQL] Replace IllegalStateException by SparkException.internalError in catalyst
• [SPARK-45762] [SC-148486][CORE] Support shuffle managers defined in user jars by changing startup order
• [SPARK-46239] [SC-149752][CORE] Hide Jetty info
• [SPARK-45886] [SC-148089][SQL] Output full stack trace in callSite of DataFrame context
• [SPARK-46290] [SC-150059][PYTHON] Change saveMode to a boolean flag for DataSourceWriter
• [SPARK-45899] [SC-148097][CONNECT] Set errorClass in errorInfoToThrowable
• [SPARK-45841] [SC-147657][SQL] Expose stack trace by DataFrameQueryContext
• [SPARK-45843] [SC-147721][CORE] Support killall in REST Submission API
• [SPARK-46234] [SC-149711][PYTHON] Introduce PySparkKeyError for PySpark error framework
• [SPARK-45606] [SC-147655][SQL] Release restrictions on multi-layer runtime filter
• [SPARK-45725] [SC-147006][SQL] Remove the non-default IN subquery runtime filter
• [SPARK-45694] [SC-147314][SPARK-45695][SQL] Clean up deprecated API usage View.force and ScalaNumberProxy.signum
• [SPARK-45805] [SC-147488][SQL] Make withOrigin more generic
• [SPARK-46292] [SC-150086][CORE][UI] Show a summary of workers in MasterPage
• [SPARK-45685] [SC-146881][CORE][SQL] Use LazyList instead of Stream
• [SPARK-45888] [SC-149945][SS] Apply error class framework to State (Metadata) Data Source
• [SPARK-46261] [SC-150111][CONNECT] DataFrame.withColumnsRenamed should keep the dict/map ordering
• [SPARK-46263] [SC-149933][SQL][SS][ML][MLLIB][UI] Clean up SeqOps.view and ArrayOps.view conversions
• [SPARK-46029] [SC-149255][SQL] Escape the single quote, _ and % for DS V2 pushdown
• [SPARK-46275] [SC-150189] Protobuf: Return null in permissive mode when deserialization fails.
• [SPARK-46320] [SC-150187][CORE] Support spark.master.rest.host
• [SPARK-46092] [SC-149658][SQL] Don’t push down Parquet row group filters that overflow
• [SPARK-46300] [SC-150097][PYTHON][CONNECT] Match minor behavior matching in Column with full test coverage
• [SPARK-46298] [SC-150092][PYTHON][CONNECT] Match deprecation warning, test case, and error of Catalog.createExternalTable
• [SPARK-45689] [SC-146898][SPARK-45690][SPARK-45691][CORE][SQL] Clean up the deprecated API usage related to StringContext/StringOps/RightProjection/LeftProjection/Either and type use of BufferedIterator/CanBuildFrom/Traversable
• [SPARK-33393] [SC-148407][SQL] Support SHOW TABLE EXTENDED in v2
• [SPARK-45737] [SC-147011][SQL] Remove unnecessary .toArray[InternalRow] in SparkPlan#executeTake
• [SPARK-46249] [SC-149797][SS] Require instance lock for acquiring RocksDB metrics to prevent race with background operations
• [SPARK-46260] [SC-149940][PYTHON][SQL] DataFrame.withColumnsRenamed should respect the dict ordering
• [SPARK-46274] [SC-150043] [SQL] Fix Range operator computeStats() to check long validity before converting
• [SPARK-46040] [SC-149767][SQL][Python] Update UDTF API for ‘analyze’ partitioning/ordering columns to support general expressions
• [SPARK-46287] [SC-149949][PYTHON][CONNECT] DataFrame.isEmpty should work with all datatypes
• [SPARK-45678] [SC-146825][CORE] Cover BufferReleasingInputStream.available/reset under tryOrFetchFailedException
• [SPARK-45667] [SC-146716][CORE][SQL][CONNECT] Clean up the deprecated API usage related to IterableOnceExtensionMethods.
• [SPARK-43980] [SC-148992][SQL] introducing select * except syntax
• [SPARK-46269] [SC-149816][PS] Enable more NumPy compatibility function tests
• [SPARK-45807] [SC-149851][SQL] Add createOrReplaceView(..) / replaceView(..) to ViewCatalog
• [SPARK-45742] [SC-147212][CORE][CONNECT][MLLIB][PYTHON] Introduce an implicit function for Scala Array to wrap into immutable.ArraySeq.
• [SPARK-46205] [SC-149655][CORE] Improve PersistenceEngine performance with KryoSerializer
• [SPARK-45746] [SC-149359][Python] Return specific error messages if UDTF ‘analyze’ or ‘eval’ method accepts or returns wrong values
• [SPARK-46186] [SC-149707][CONNECT] Fix illegal state transition when ExecuteThreadRunner interrupted before started
• [SPARK-46262] [SC-149803][PS] Enable test for np.left_shift for Pandas-on-Spark object.
• [SPARK-45093] [SC-149679][CONNECT][PYTHON] Properly support error handling and conversion for AddArtifactHandler
• [SPARK-46188] [SC-149571][Doc][3.5] Fix the CSS of Spark doc’s generated tables
• [SPARK-45940] [SC-149549][PYTHON] Add InputPartition to DataSourceReader interface
• [SPARK-43393] [SC-148301][SQL] Address sequence expression overflow bug.
• [SPARK-46144] [SC-149466][SQL] Fail INSERT INTO … REPLACE statement if the condition contains subquery
• [SPARK-46118] [SC-149096][SQL][SS][CONNECT] Use SparkSession.sessionState.conf instead of SQLContext.conf and mark SQLContext.conf as deprecated
• [SPARK-45760] [SC-147550][SQL] Add With expression to avoid duplicating expressions
• [SPARK-43228] [SC-149670][SQL] Join keys also match PartitioningCollection in CoalesceBucketsInJoin
• [SPARK-46223] [SC-149691][PS] Test SparkPandasNotImplementedError with cleaning up unused code
• [SPARK-46119] [SC-149205][SQL] Override toString method for UnresolvedAlias
• [SPARK-46206] [SC-149674][PS] Use a narrower scope exception for SQL processor
• [SPARK-46191] [SC-149565][CORE] Improve FileSystemPersistenceEngine.persist error msg in case of the existing file
• [SPARK-46036] [SC-149356][SQL] Removing error-class from raise_error function
• [SPARK-46055] [SC-149254][SQL] Rewrite Catalog Database APIs implementation
• [SPARK-46152] [SC-149454][SQL] XML: Add DecimalType support in XML schema inference
• [SPARK-45943] [SC-149452][SQL] Move DetermineTableStats to resolution rules
• [SPARK-45887] [SC-148994][SQL] Align codegen and non-codegen implementation of Encode
• [SPARK-45022] [SC-147158][SQL] Provide context for dataset API errors
• [SPARK-45469] [SC-145135][CORE][SQL][CONNECT][PYTHON] Replace toIterator with iterator for IterableOnce
• [SPARK-46141] [SC-149357][SQL] Change default for spark.sql.legacy.ctePrecedencePolicy to CORRECTED
• [SPARK-45663] [SC-146617][CORE][MLLIB] Replace IterableOnceOps#aggregate with IterableOnceOps#foldLeft
• [SPARK-45660] [SC-146493][SQL] Re-use Literal objects in ComputeCurrentTime rule
• [SPARK-45803] [SC-147489][CORE] Remove the no longer used RpcAbortException
• [SPARK-46148] [SC-149350][PS] Fix pyspark.pandas.mlflow.load_model test (Python 3.12)
• [SPARK-46110] [SC-149090][PYTHON] Use error classes in catalog, conf, connect, observation, pandas modules
• [SPARK-45827] [SC-149203][SQL] Variant fixes with codegen and vectorized reader disabled
• [SPARK-46080] Revert “[SC-149012][PYTHON] Upgrade Cloudpickle to 3.0.0”
• [SPARK-45460] [SC-144852][SQL] Replace scala.collection.convert.ImplicitConversions to scala.jdk.CollectionConverters
• [SPARK-46073] [SC-149019][SQL] Remove the special resolution of UnresolvedNamespace for certain commands
• [SPARK-45600] [SC-148976][PYTHON] Make Python data source registration session level
• [SPARK-46074] [SC-149063][CONNECT][SCALA] Insufficient details in error message on UDF failure
• [SPARK-46114] [SC-149076][PYTHON] Add PySparkIndexError for error framework
• [SPARK-46120] [SC-149083][CONNECT][PYTHON] Remove helper function DataFrame.withPlan
• [SPARK-45927] [SC-148660][PYTHON] Update path handling for Python data source
• [SPARK-46062] [14.x][SC-148991][SQL] Sync the isStreaming flag between CTE definition and reference
• [SPARK-45698] [SC-146736][CORE][SQL][SS] Clean up the deprecated API usage related to Buffer
• [SPARK-46064] [SC-148984][SQL][SS] Move out EliminateEventTimeWatermark to the analyzer and change to only take effect on resolved child
• [SPARK-45922] [SC-149038][CONNECT][CLIENT] Minor retries refactoring (follow-up to multiple policies)
• [SPARK-45851] [SC-148419][CONNECT][SCALA] Support multiple policies in scala client
• [SPARK-45974] [SC-149062][SQL] Add scan.filterAttributes non-empty judgment for RowLevelOperationRuntimeGroupFiltering
• [SPARK-46085] [SC-149049][CONNECT] Dataset.groupingSets in Scala Spark Connect client
• [SPARK-45136] [SC-146607][CONNECT] Enhance ClosureCleaner with Ammonite support
• [SPARK-46084] [SC-149020][PS] Refactor data type casting operation for Categorical type.
• [SPARK-46083] [SC-149018][PYTHON] Make SparkNoSuchElementException as a canonical error API
• [SPARK-46048] [SC-148982][PYTHON][CONNECT] Support DataFrame.groupingSets in Python Spark Connect
• [SPARK-46080] [SC-149012][PYTHON] Upgrade Cloudpickle to 3.0.0
• [SPARK-46082] [SC-149015][PYTHON][CONNECT] Fix protobuf string representation for Pandas Functions API with Spark Connect
• [SPARK-46065] [SC-148985][PS] Refactor (DataFrame|Series).factorize() to use create_map.
• [SPARK-46070] [SC-148993][SQL] Compile regex pattern in SparkDateTimeUtils.getZoneId outside the hot loop
• [SPARK-46063] [SC-148981][PYTHON][CONNECT] Improve error messages related to argument types in cute, rollup, groupby, and pivot
• [SPARK-44442] [SC-144120][MESOS] Remove Mesos support
• [SPARK-45555] [SC-147620][PYTHON] Includes a debuggable object for failed assertion
• [SPARK-46048] [SC-148871][PYTHON][SQL] Support DataFrame.groupingSets in PySpark
• [SPARK-45768] [SC-147304][SQL][PYTHON] Make faulthandler a runtime configuration for Python execution in SQL
• [SPARK-45996] [SC-148586][PYTHON][CONNECT] Show proper dependency requirement messages for Spark Connect
• [SPARK-45681] [SC-146893][UI] Clone a js version of UIUtils.errorMessageCell for consistent error rendering on UI
• [SPARK-45767] [SC-147225][CORE] Delete TimeStampedHashMap and its UT
• [SPARK-45696] [SC-148864][CORE] Fix method tryCompleteWith in trait Promise is deprecated
• [SPARK-45503] [SC-146338][SS] Add Conf to Set RocksDB Compression
• [SPARK-45338] [SC-143934][CORE][SQL] Replace scala.collection.JavaConverters to scala.jdk.CollectionConverters
• [SPARK-44973] [SC-148804][SQL] Fix ArrayIndexOutOfBoundsException in conv()
• [SPARK-45912] [SC-144982][SC-148467][SQL] Enhancement of XSDToSchema API: Change to HDFS API for cloud storage accessibility
• [SPARK-45819] [SC-147570][CORE] Support clear in REST Submission API
• [SPARK-45552] [14.x][SC-146874][PS] Introduce flexible parameters to assertDataFrameEqual
• [SPARK-45815] [SC-147877][SQL][Streaming][14.x][CHERRY-PICK] Provide an interface for other Streaming sources to add _metadata columns
• [SPARK-45929] [SC-148692][SQL] Support groupingSets operation in dataframe api
• [SPARK-46012] [SC-148693][CORE] EventLogFileReader should not read rolling logs if app status file is missing
• [SPARK-45965] [SC-148575][SQL] Move DSv2 partitioning expressions into functions.partitioning
• [SPARK-45971] [SC-148541][CORE][SQL] Correct the package name of SparkCollectionUtils to org.apache.spark.util
• [SPARK-45688] [SC-147356][SPARK-45693][CORE] Clean up the deprecated API usage related to MapOps & Fix method += in trait Growable is deprecated
• [SPARK-45962] [SC-144645][SC-148497][SQL] Remove treatEmptyValuesAsNulls and use nullValue option instead in XML
• [SPARK-45988] [SC-148580][SPARK-45989][PYTHON] Fix typehints to handle list GenericAlias in Python 3.11+
• [SPARK-45999] [SC-148595][PS] Use dedicated PandasProduct in cumprod
• [SPARK-45986] [SC-148583][ML][PYTHON] Fix pyspark.ml.torch.tests.test_distributor in Python 3.11
• [SPARK-45994] [SC-148581][PYTHON] Change description-file to description_file
• [SPARK-45575] [SC-146713][SQL] Support time travel options for df read API
• [SPARK-45747] [SC-148404][SS] Use prefix key information in state metadata to handle reading state for session window aggregation
• [SPARK-45898] [SC-148213][SQL] Rewrite catalog table APIs to use unresolved logical plan
• [SPARK-45828] [SC-147663][SQL] Remove deprecated method in dsl
• [SPARK-45990] [SC-148576][SPARK-45987][PYTHON][CONNECT] Upgrade protobuf to 4.25.1 to support Python 3.11
• [SPARK-45911] [SC-148226][CORE] Make TLS1.3 the default for RPC SSL
• [SPARK-45955] [SC-148429][UI] Collapse Support for Flamegraph and thread dump details
• [SPARK-45936] [SC-148392][PS] Optimize Index.symmetric_difference
• [SPARK-45389] [SC-144834][SQL][3.5] Correct MetaException matching rule on getting partition metadata
• [SPARK-45655] [SC-148085][SQL][SS] Allow non-deterministic expressions inside AggregateFunctions in CollectMetrics
• [SPARK-45946] [SC-148448][SS] Fix use of deprecated FileUtils write to pass default charset in RocksDBSuite
• [SPARK-45511] [SC-148285][SS] State Data Source - Reader
• [SPARK-45918] [SC-148293][PS] Optimize MultiIndex.symmetric_difference
• [SPARK-45813] [SC-148288][CONNECT][PYTHON] Return the observed metrics from commands
• [SPARK-45867] [SC-147916][CORE] Support spark.worker.idPattern
• [SPARK-45945] [SC-148418][CONNECT] Add a helper function for parser
• [SPARK-45930] [SC-148399][SQL] Support non-deterministic UDFs in MapInPandas/MapInArrow
• [SPARK-45810] [SC-148356][Python] Create Python UDTF API to stop consuming rows from the input table
• [SPARK-45731] [SC-147929][SQL] Also update partition statistics with ANALYZE TABLE command
• [SPARK-45868] [SC-148355][CONNECT] Make sure spark.table use the same parser with vanilla spark
• [SPARK-45882] [SC-148161][SQL] BroadcastHashJoinExec propagate partitioning should respect CoalescedHashPartitioning
• [SPARK-45710] [SC-147428][SQL] Assign names to error _LEGACY_ERROR_TEMP_21[59,60,61,62]
• [SPARK-45915] [SC-148306][SQL] Treat decimal(x, 0) the same as IntegralType in PromoteStrings
• [SPARK-45786] [SC-147552][SQL] Fix inaccurate Decimal multiplication and division results
• [SPARK-45871] [SC-148084][CONNECT] Optimizations collection conversion related to .toBuffer in the connect modules
• [SPARK-45822] [SC-147546][CONNECT] SparkConnectSessionManager may look up a stopped sparkcontext
• [SPARK-45913] [SC-148289][PYTHON] Make the internal attributes private from PySpark errors.
• [SPARK-45827] [SC-148201][SQL] Add Variant data type in Spark.
• [SPARK-44886] [SC-147876][SQL] Introduce CLUSTER BY clause for CREATE/REPLACE TABLE
• [SPARK-45756] [SC-147959][CORE] Support spark.master.useAppNameAsAppId.enabled
• [SPARK-45798] [SC-147956][CONNECT] Assert server-side session ID
• [SPARK-45896] [SC-148091][SQL] Construct ValidateExternalType with the correct expected type
• [SPARK-45902] [SC-148123][SQL] Remove unused function resolvePartitionColumns from DataSource
• [SPARK-45909] [SC-148137][SQL] Remove NumericType cast if it can safely up-cast in IsNotNull
• [SPARK-42821] [SC-147715][SQL] Remove unused parameters in splitFiles methods
• [SPARK-45875] [SC-148088][CORE] Remove MissingStageTableRowData from core module
• [SPARK-45783] [SC-148120][PYTHON][CONNECT] Improve error messages when Spark Connect mode is enabled but remote URL is not set
• [SPARK-45804] [SC-147529][UI] Add spark.ui.threadDump.flamegraphEnabled config to switch flame graph on/off
• [SPARK-45664] [SC-146878][SQL] Introduce a mapper for orc compression codecs
• [SPARK-45481] [SC-146715][SQL] Introduce a mapper for parquet compression codecs
• [SPARK-45752] [SC-148005][SQL] Simplify the code for check unreferenced CTE relations
• [SPARK-44752] [SC-146262][SQL] XML: Update Spark Docs
• [SPARK-45752] [SC-147869][SQL] Unreferenced CTE should all be checked by CheckAnalysis0
• [SPARK-45842] [SC-147853][SQL] Refactor Catalog Function APIs to use analyzer
• [SPARK-45639] [SC-147676][SQL][PYTHON] Support loading Python data sources in DataFrameReader
• [SPARK-45592] [SC-146977][SQL] Correctness issue in AQE with InMemoryTableScanExec
• [SPARK-45794] [SC-147646][SS] Introduce state metadata source to query the streaming state metadata information
• [SPARK-45739] [SC-147059][PYTHON] Catch IOException instead of EOFException alone for faulthandler
• [SPARK-45677] [SC-146700] Better error logging for Observation API
• [SPARK-45544] [SC-146873][CORE] Integrate SSL support into TransportContext
• [SPARK-45654] [SC-147152][PYTHON] Add Python data source write API
• [SPARK-45771] [SC-147270][CORE] Enable spark.eventLog.rolling.enabled by default
• [SPARK-45713] [SC-146986][PYTHON] Support registering Python data sources
• [SPARK-45656] [SC-146488][SQL] Fix observation when named observations with the same name on different datasets
• [SPARK-45808] [SC-147571][CONNECT][PYTHON] Better error handling for SQL Exceptions
• [SPARK-45774] [SC-147353][CORE][UI] Support spark.master.ui.historyServerUrl in ApplicationPage
• [SPARK-45776] [SC-147327][CORE] Remove the defensive null check for MapOutputTrackerMaster#unregisterShuffle added in SPARK-39553
• [SPARK-45780] [SC-147410][CONNECT] Propagate all Spark Connect client threadlocals in InheritableThread
• [SPARK-45785] [SC-147419][CORE] Support spark.deploy.appNumberModulo to rotate app number
• [SPARK-45793] [SC-147456][CORE] Improve the built-in compression codecs
• [SPARK-45757] [SC-147282][ML] Avoid re-computation of NNZ in Binarizer
• [SPARK-45209] [SC-146490][CORE][UI] Flame Graph Support For Executor Thread Dump Page
• [SPARK-45777] [SC-147319][CORE] Support spark.test.appId in LocalSchedulerBackend
• [SPARK-45523] [SC-146331][PYTHON] Refactor the null-checking to have shortcuts
• [SPARK-45711] [SC-146854][SQL] Introduce a mapper for avro compression codecs
• [SPARK-45523] [SC-146077][Python] Return useful error message if UDTF returns None for any non-nullable column
• [SPARK-45614] [SC-146729][SQL] Assign names to error _LEGACY_ERROR_TEMP_215[6,7,8]
• [SPARK-45755] [SC-147164][SQL] Improve Dataset.isEmpty() by applying global limit 1
• [SPARK-45569] [SC-145915][SQL] Assign name to the error _LEGACY_ERROR_TEMP_2153
• [SPARK-45749] [SC-147153][CORE][WEBUI] Fix Spark History Server to sort Duration column properly
• [SPARK-45754] [SC-147169][CORE] Support spark.deploy.appIdPattern
• [SPARK-45707] [SC-146880][SQL] Simplify DataFrameStatFunctions.countMinSketch with CountMinSketchAgg
• [SPARK-45753] [SC-147167][CORE] Support spark.deploy.driverIdPattern
• [SPARK-45763] [SC-147177][CORE][UI] Improve MasterPage to show Resource column only when it exists
• [SPARK-45112] [SC-143259][SQL] Use UnresolvedFunction based resolution in SQL Dataset functions

Databricks ODBC/JDBC driver support

Databricks supports ODBC/JDBC drivers released in the past 2 years. Please download the recently released drivers and upgrade (download ODBC, download JDBC).

Maintenance updates

See Databricks Runtime 14.3 maintenance updates. This is commented out at initial release.

System environment

• Operating System: Ubuntu 22.04.3 LTS
• Java: Zulu 8.74.0.17-CA-linux64
• Scala: 2.12.15
• Python: 3.10.12
• R: 4.3.1
• Delta Lake: 3.1.0

Installed Python libraries

Installed R libraries

R libraries are installed from the Posit Package Manager CRAN snapshot on 2023-02-10.

Installed Java and Scala libraries (Scala 2.12 cluster version)