Databricks Runtime 11.2 (EoS)

The following release notes provide information about Databricks Runtime 11.2, powered by Apache Spark 3.3.0. Databricks released this version in September 2022.

New features and improvements

• Change to ALTER TABLE permissions in table access controls (table ACLs)
• Dynamic pruning for MERGE INTO
• Improved conflict detection in Delta with dynamic file pruning
• New Open Source Integrations card in DSE/SQL homepages
• CONVERT TO DELTA partition detection improvements
• Table schemas now support default values for columns
• New H3 geospatial functions
• New Databricks Runtime dependency
• Bring your own key: Git credentials encryption
• SQL: New aggregate function any_value
• Databricks Utilities file system commands allowed on more cluster types
• CREATE privileges can now be granted on metastores for Unity Catalog
• Optimized writes for unpartitioned tables in Photon-enabled clusters
• Photon support for more data sources
• SQL: ALTER SHARE now supports START VERSION

Change to ALTER TABLE permissions in table access controls (table ACLs)

Now, users only need MODIFY permissions to change a table’s schema or properties with ALTER TABLE. Ownership is still required to grant permissions on a table, change its owner and location, or rename it. This change makes the permission model for table ACLs consistent with Unity Catalog. See ALTER TABLE.

Dynamic pruning for MERGE INTO

When using Photon-enabled compute, MERGE INTO now uses dynamic file and partition pruning when appropriate to improve performance, for example when a small source table is merged into a larger target table.

Improved conflict detection in Delta with dynamic file pruning

When checking for potential conflicts during commits, conflict detection now considers files that are pruned by dynamic file pruning, but would not have been pruned by static filters. This results in a decreased number of failed transactions.

New Open Source Integrations card in DSE/SQL homepages

Introducing the new ‘Open Source Integrations’ card in the DSE/SQL homepages that displays open source integration options such as Delta Live Tables and dbt core.

CONVERT TO DELTA partition detection improvements

CONVERT TO DELTA automatically infers the partition schema for Parquet tables registered to the Hive metastore or Unity Catalog, eliminating the need to provide the PARTITIONED BY clause.

CONVERT TO DELTA leverages partition information from the metastore to discover files for a Parquet table instead of listing the entire base path, ensuring dropped partitions do not get added to the Delta table.

See Convert to Delta Lake.

Table schemas now support default values for columns

Table schemas now support setting default values for columns. INSERT, UPDATE, and DELETE commands for these columns can refer to these values with the DEFAULT keyword. For example, CREATE TABLE t (id INT, data INT DEFAULT 42) USING PARQUET followed by INSERT INTO t VALUES (1, DEFAULT) will append the row (1, 42). This behavior is supported for CSV, JSON, Orc, and Parquet data sources.

New H3 geospatial functions

You can now use 28 new built-in H3 expressions for geospatial processing in Photon-enabled clusters, available in SQL, Scala, and Python. See H3 geospatial functions.

New Databricks Runtime dependency

Databricks Runtime now depends on the H3 Java library version 3.7.0.

Bring your own key: Git credentials encryption

You can use Azure Key Vault to encrypt a Git personal access token (PAT) or other Git credential.

See Set up Databricks Git folders (Repos).

SQL: New aggregate function any_value

The new any_value aggregate function returns any random value of expr for a group of rows. See the any_value aggregate function.

Databricks Utilities file system commands allowed on more cluster types

dbutils.fs commands (except for mount-related commands) are now allowed on user-isolation clusters with Unity Catalog, as well as legacy table ACL clusters when the user has ANY FILE permissions.

CREATE privileges can now be granted on metastores for Unity Catalog

CREATE CATALOG, CREATE EXTERNAL LOCATION, CREATE SHARE, CREATE RECIPIENT, and CREATE PROVIDER privileges can now be granted on Unity Catalog metastores.

Optimized writes for unpartitioned tables in Photon-enabled clusters

Unity Catalog managed tables now automatically persist files of a well-tuned size from unpartitioned tables to improve query speed and optimize performance.

Photon support for more data sources

Photon now supports more data sources, including CSV and Avro, and is also compatible with cached data frames. Previously, scanning these data sources meant that the entire query could not be photonized, regardless of the query’s operators or expressions. Now, queries scanning these data sources can be photonized, unlocking significant latency and TCO improvements.

This feature is enabled by default by the spark.databricks.photon.photonRowToColumnar.enabled configuration.

Limitations:

• Schemas with nested types are unsupported (i.e arrays, maps, and structs) in this release.
• ORC, RDD, Kinesis, Kafka, and EventHub sources are unsupported in this release.

SQL: ALTER SHARE now supports START VERSION

The ALTER SHARE command now supports START VERSION, which which allows providers to share data starting from a specific table version. See ALTER SHARE.

Library upgrades

• Upgraded Python libraries:
  • distlib from 0.3.4 to 0.3.5
  • filelock from 3.7.1 to 3.8.0
  • plotly from 5.8.2 to 5.9.0
  • protobuf from 4.21.2 to 4.21.5
• Upgraded R libraries:
  • broom from 0.8.0 to 1.0.0
  • bslib from 0.3.1 to 0.4.0
  • callr from 3.7.0 to 3.7.1
  • caret from 6.0-92 to 6.0-93
  • dbplyr from 2.2.0 to 2.2.1
  • devtools from 2.4.3 to 2.4.4
  • evaluate from 0.15 to 0.16
  • farver from 2.1.0 to 2.1.1
  • fontawesome from 0.2.2 to 0.3.0
  • future from 1.26.1 to 1.27.0
  • generics from 0.1.2 to 0.1.3
  • gert from 1.6.0 to 1.7.0
  • globals from 0.15.1 to 0.16.0
  • googlesheets4 from 1.0.0 to 1.0.1
  • hardhat from 1.1.0 to 1.2.0
  • htmltools from 0.5.2 to 0.5.3
  • parallelly from 1.32.0 to 1.32.1
  • pillar from 1.7.0 to 1.8.0
  • pkgload from 1.2.4 to 1.3.0
  • processx from 3.6.1 to 3.7.0
  • Rcpp from 1.0.8.3 to 1.0.9
  • recipes from 0.2.0 to 1.0.1
  • rlang from 1.0.2 to 1.0.4
  • roxygen2 from 7.2.0 to 7.2.1
  • RSQLite from 2.2.14 to 2.2.15
  • sass from 0.4.1 to 0.4.2
  • shiny from 1.7.1 to 1.7.2
  • stringi from 1.7.6 to 1.7.8
  • tibble from 3.1.7 to 3.1.8
  • tidyverse from 1.3.1 to 1.3.2
  • timeDate from 3043.102 to 4021.104
  • xfun from 0.31 to 0.32
• Upgraded Java libraries:
  • org.apache.orc.orc-core from 1.7.4 to 1.7.5
  • org.apache.orc.orc-mapreduce from 1.7.4 to 1.7.5
  • org.apache.orc.orc-shims from 1.7.4 to 1.7.5

Apache Spark

Databricks Runtime 11.2 includes Apache Spark 3.3.0. This release includes all Spark fixes and improvements included in Databricks Runtime 11.1 (EoS), as well as the following additional bug fixes and improvements made to Spark:

• [SPARK-40151] [WARMFIX][SC-109002][SC-108809][SQL] Return wider ANSI interval types from the percentile functions
• [SPARK-40054] [SQL] Restore the error handling syntax of try_cast()
• [SPARK-39489] [CORE] Improve event logging JsonProtocol performance by using Jackson instead of Json4s
• [SPARK-39319] [CORE][SQL] Make query contexts as a part of SparkThrowable
• [SPARK-40085] [SQL] Use INTERNAL_ERROR error class instead of IllegalStateException to indicate bugs
• [SPARK-40001] [SQL] Make NULL writes to JSON DEFAULT columns write ‘null’ to storage
• [SPARK-39635] [SQL] Support driver metrics in DS v2 custom metric API
• [SPARK-39184] [SQL] Handle undersized result array in date and timestamp sequences
• [SPARK-40019] [SQL] Refactor comment of ArrayType’s containsNull and refactor the misunderstanding logics in collectionOperator’s expression about containsNull
• [SPARK-39989] [SQL] Support estimate column statistics if it is foldable expression
• [SPARK-39926] [SQL] Fix bug in column DEFAULT support for non-vectorized Parquet scans
• [SPARK-40052] [SQL] Handle direct byte buffers in VectorizedDeltaBinaryPackedReader
• [SPARK-40044] [SQL] Fix the target interval type in cast overflow errors
• [SPARK-39835] [SQL] Fix EliminateSorts remove global sort below the local sort
• [SPARK-40002] [SQL] Don’t push down limit through window using ntile
• [SPARK-39976] [SQL] ArrayIntersect should handle null in left expression correctly
• [SPARK-39985] [SQL] Enable implicit DEFAULT column values in inserts from DataFrames
• [SPARK-39776] [SQL] JOIN verbose string should add Join type
• [SPARK-38901] [SQL] DS V2 supports push down misc functions
• [SPARK-40028] [SQL][FollowUp] Improve examples of string functions
• [SPARK-39983] [CORE][SQL] Do not cache unserialized broadcast relations on the driver
• [SPARK-39812] [SQL] Simplify code which construct AggregateExpression with toAggregateExpression
• [SPARK-40028] [SQL] Add binary examples for string expressions
• [SPARK-39981] [SQL] Throw the exception QueryExecutionErrors.castingCauseOverflowErrorInTableInsert in Cast
• [SPARK-40007] [PYTHON][SQL] Add ‘mode’ to functions
• [SPARK-40008] [SQL] Support casting of integrals to ANSI intervals
• [SPARK-40003] [PYTHON][SQL] Add ‘median’ to functions
• [SPARK-39952] [SQL] SaveIntoDataSourceCommand should recache result relation
• [SPARK-39951] [SQL] Update Parquet V2 columnar check for nested fields
• [SPARK-33236] [shuffle] Backport to DBR 11.x: Enable Push-based shuffle service to store state in NM level DB for work preserving restart
• [SPARK-39836] [SQL] Simplify V2ExpressionBuilder by extract common method.
• [SPARK-39873] [SQL] Remove OptimizeLimitZero and merge it into EliminateLimits
• [SPARK-39961] [SQL] DS V2 push-down translate Cast if the cast is safe
• [SPARK-39872] [SQL] Change to use BytePackerForLong#unpack8Values with Array input api in VectorizedDeltaBinaryPackedReader
• [SPARK-39858] [SQL] Remove unnecessary AliasHelper or PredicateHelper for some rules
• [SPARK-39900] [SQL] Address partial or negated condition in binary format’s predicate pushdown
• [SPARK-39904] [SQL] Rename inferDate to prefersDate and clarify semantics of the option in CSV data source
• [SPARK-39958] [SQL] Add warning log when unable to load custom metric object
• [SPARK-39932] [SQL] WindowExec should clear the final partition buffer
• [SPARK-37194] [SQL] Avoid unnecessary sort in v1 write if it’s not dynamic partition
• [SPARK-39902] [SQL] Add Scan details to spark plan scan node in SparkUI
• [SPARK-39865] [SQL] Show proper error messages on the overflow errors of table insert
• [SPARK-39940] [SS] Refresh catalog table on streaming query with DSv1 sink
• [SPARK-39827] [SQL] Use the error class ARITHMETIC_OVERFLOW on int overflow in add_months()
• [SPARK-39914] [SQL] Add DS V2 Filter to V1 Filter conversion
• [SPARK-39857] [SQL] Manual DBR 11.x backport; V2ExpressionBuilder uses the wrong LiteralValue data type for In predicate #43454
• [SPARK-39840] [SQL][PYTHON] Factor PythonArrowInput out as a symmetry to PythonArrowOutput
• [SPARK-39651] [SQL] Prune filter condition if compare with rand is deterministic
• [SPARK-39877] [PYTHON] Add unpivot to PySpark DataFrame API
• [SPARK-39909] [SQL] Organize the check of push down information for JDBCV2Suite
• [SPARK-39834] [SQL][SS] Include the origin stats and constraints for LogicalRDD if it comes from DataFrame
• [SPARK-39849] [SQL] Dataset.as(StructType) fills missing new columns with null value
• [SPARK-39860] [SQL] More expressions should extend Predicate
• [SPARK-39823] [SQL][PYTHON] Rename Dataset.as as Dataset.to and add DataFrame.to in PySpark
• [SPARK-39918] [SQL][MINOR] Replace the wording “un-comparable” with “incomparable” in error message
• [SPARK-39857] [SQL][3.3] V2ExpressionBuilder uses the wrong LiteralValue data type for In predicate
• [SPARK-39862] [SQL] Manual backport for PR 43654 targeting DBR 11.x: Update SQLConf.DEFAULT_COLUMN_ALLOWED_PROVIDERS to allow/deny ALTER TABLE … ADD COLUMN commands separately.
• [SPARK-39844] [SQL] Manual backport for PR 43652 targeting DBR 11.x
• [SPARK-39899] [SQL] Fix passing of message parameters to InvalidUDFClassException
• [SPARK-39890] [SQL] Make TakeOrderedAndProjectExec inherit AliasAwareOutputOrdering
• [SPARK-39809] [PYTHON] Support CharType in PySpark
• [SPARK-38864] [SQL] Add unpivot / melt to Dataset
• [SPARK-39864] [SQL] Lazily register ExecutionListenerBus
• [SPARK-39808] [SQL] Support aggregate function MODE
• [SPARK-39875] [SQL] Change protected method in final class to private or package-visible
• [SPARK-39731] [SQL] Fix issue in CSV and JSON data sources when parsing dates in “yyyyMMdd” format with CORRECTED time parser policy
• [SPARK-39805] [SS] Deprecate Trigger.Once and Promote Trigger.AvailableNow
• [SPARK-39784] [SQL] Put Literal values on the right side of the data source filter after translating Catalyst Expression to data source filter
• [SPARK-39672] [SQL][3.1] Fix removing project before filter with correlated subquery
• [SPARK-39552] [SQL] Unify v1 and v2 DESCRIBE TABLE
• [SPARK-39810] [SQL] Catalog.tableExists should handle nested namespace
• [SPARK-37287] [SQL] Pull out dynamic partition and bucket sort from FileFormatWriter
• [SPARK-39469] [SQL] Infer date type for CSV schema inference
• [SPARK-39148] [SQL] DS V2 aggregate push down can work with OFFSET or LIMIT
• [SPARK-39818] [SQL] Fix bug in ARRAY, STRUCT, MAP types with DEFAULT values with NULL field(s)
• [SPARK-39792] [SQL] Add DecimalDivideWithOverflowCheck for decimal average
• [SPARK-39798] [SQL] Replcace toSeq.toArray with .toArray[Any] in constructor of GenericArrayData
• [SPARK-39759] [SQL] Implement listIndexes in JDBC (H2 dialect)
• [SPARK-39385] [SQL] Supports push down REGR_AVGX and REGR_AVGY
• [SPARK-39787] [SQL] Use error class in the parsing error of function to_timestamp
• [SPARK-39760] [PYTHON] Support Varchar in PySpark
• [SPARK-39557] [SQL] Manual backport to DBR 11.x: Support ARRAY, STRUCT, MAP types as DEFAULT values
• [SPARK-39758] [SQL][3.3] Fix NPE from the regexp functions on invalid patterns
• [SPARK-39749] [SQL] ANSI SQL mode: Use plain string representation on casting Decimal to String
• [SPARK-39704] [SQL] Implement createIndex & dropIndex & indexExists in JDBC (H2 dialect)
• [SPARK-39803] [SQL] Use LevenshteinDistance instead of StringUtils.getLevenshteinDistance
• [SPARK-39339] [SQL] Support TimestampNTZ type in JDBC data source
• [SPARK-39781] [SS] Add support for providing max_open_files to rocksdb state store provider
• [SPARK-39719] [R] Implement databaseExists/getDatabase in SparkR support 3L namespace
• [SPARK-39751] [SQL] Rename hash aggregate key probes metric
• [SPARK-39772] [SQL] namespace should be null when database is null in the old constructors
• [SPARK-39625] [SPARK-38904][SQL] Add Dataset.as(StructType)
• [SPARK-39384] [SQL] Compile built-in linear regression aggregate functions for JDBC dialect
• [SPARK-39720] [R] Implement tableExists/getTable in SparkR for 3L namespace
• [SPARK-39744] [SQL] Add the REGEXP_INSTR function
• [SPARK-39716] [R] Make currentDatabase/setCurrentDatabase/listCatalogs in SparkR support 3L namespace
• [SPARK-39788] [SQL] Rename catalogName to dialectName for JdbcUtils
• [SPARK-39647] [CORE] Register the executor with ESS before registering the BlockManager
• [SPARK-39754] [CORE][SQL] Remove unused import or unnecessary {}
• [SPARK-39706] [SQL] Set missing column with defaultValue as constant in ParquetColumnVector
• [SPARK-39699] [SQL] Make CollapseProject smarter about collection creation expressions
• [SPARK-39737] [SQL] PERCENTILE_CONT and PERCENTILE_DISC should support aggregate filter
• [SPARK-39579] [SQL][PYTHON][R] Make ListFunctions/getFunction/functionExists compatible with 3 layer namespace
• [SPARK-39627] [SQL] JDBC V2 pushdown should unify the compile API
• [SPARK-39748] [SQL][SS] Include the origin logical plan for LogicalRDD if it comes from DataFrame
• [SPARK-39385] [SQL] Translate linear regression aggregate functions for pushdown
• [SPARK-39695] [SQL] Add the REGEXP_SUBSTR function
• [SPARK-39667] [SQL] Add another workaround when there is not enough memory to build and broadcast the table
• [SPARK-39666] [ES-337834][SQL] Use UnsafeProjection.create to respect spark.sql.codegen.factoryMode in ExpressionEncoder
• [SPARK-39643] [SQL] Prohibit subquery expressions in DEFAULT values
• [SPARK-38647] [SQL] Add SupportsReportOrdering mix in interface for Scan (DataSourceV2)
• [SPARK-39497] [SQL] Improve the analysis exception of missing map key column
• [SPARK-39661] [SQL] Avoid creating unnecessary SLF4J Logger
• [SPARK-39713] [SQL] ANSI mode: add suggestion of using try_element_at for INVALID_ARRAY_INDEX error
• [SPARK-38899] [SQL]DS V2 supports push down datetime functions
• [SPARK-39638] [SQL] Change to use ConstantColumnVector to store partition columns in OrcColumnarBatchReader
• [SPARK-39653] [SQL] Clean up ColumnVectorUtils#populate(WritableColumnVector, InternalRow, int) from ColumnVectorUtils
• [SPARK-39231] [SQL] Use ConstantColumnVector instead of On/OffHeapColumnVector to store partition columns in VectorizedParquetRecordReader
• [SPARK-39547] [SQL] V2SessionCatalog should not throw NoSuchDatabaseException in loadNamspaceMetadata
• [SPARK-39447] [SQL] Avoid AssertionError in AdaptiveSparkPlanExec.doExecuteBroadcast
• [SPARK-39492] [SQL] Rework MISSING_COLUMN
• [SPARK-39679] [SQL] TakeOrderedAndProjectExec should respect child output ordering
• [SPARK-39606] [SQL] Use child stats to estimate order operator
• [SPARK-39611] [PYTHON][PS] Fix wrong aliases in array_ufunc
• [SPARK-39656] [SQL][3.3] Fix wrong namespace in DescribeNamespaceExec
• [SPARK-39675] [SQL] Switch ‘spark.sql.codegen.factoryMode’ configuration from testing purpose to internal purpose
• [SPARK-39139] [SQL] DS V2 supports push down DS V2 UDF
• [SPARK-39434] [SQL] Provide runtime error query context when array index is out of bounding
• [SPARK-39479] [SQL] DS V2 supports push down math functions(non ANSI)
• [SPARK-39618] [SQL] Add the REGEXP_COUNT function
• [SPARK-39553] [CORE] Multi-thread unregister shuffle shouldn’t throw NPE when using Scala 2.13
• [SPARK-38755] [PYTHON][3.3] Add file to address missing pandas general functions
• [SPARK-39444] [SQL] Add OptimizeSubqueries into nonExcludableRules list
• [SPARK-39316] [SQL] Merge PromotePrecision and CheckOverflow into decimal binary arithmetic
• [SPARK-39505] [UI] Escape log content rendered in UI
• [SPARK-39448] [SQL] Add ReplaceCTERefWithRepartition into nonExcludableRules list
• [SPARK-37961] [SQL] Override maxRows/maxRowsPerPartition for some logical operators
• [SPARK-35223] Revert Add IssueNavigationLink
• [SPARK-39633] [SQL] Support timestamp in seconds for TimeTravel using Dataframe options
• [SPARK-38796] [SQL] Update documentation for number format strings with the {try_}to_number functions
• [SPARK-39650] [SS] Fix incorrect value schema in streaming deduplication with backward compatibility
• [SPARK-39636] [CORE][UI] Fix multiple bugs in JsonProtocol, impacting off heap StorageLevels and Task/Executor ResourceRequests
• [SPARK-39432] [SQL] Return ELEMENT_AT_BY_INDEX_ZERO from element_at(*, 0)
• [SPARK-39349] Add a centralized CheckError method for QA of error path
• [SPARK-39453] [SQL] DS V2 supports push down misc non-aggregate functions(non ANSI)
• [SPARK-38978] [SQL] DS V2 supports push down OFFSET operator
• [SPARK-39567] [SQL] Support ANSI intervals in the percentile functions
• [SPARK-39383] [SQL] Support DEFAULT columns in ALTER TABLE ALTER COLUMNS to V2 data sources
• [SPARK-39396] [SQL] Fix LDAP login exception ‘error code 49 - invalid credentials’
• [SPARK-39548] [SQL] CreateView Command with a window clause query hit a wrong window definition not found issue
• [SPARK-39575] [AVRO] add ByteBuffer#rewind after ByteBuffer#get in Avr…
• [SPARK-39543] The option of DataFrameWriterV2 should be passed to storage properties if fallback to v1
• [SPARK-39564] [SS] Expose the information of catalog table to the logical plan in streaming query
• [SPARK-39582] [SQL] Fix “Since” marker for array_agg
• [SPARK-39388] [SQL] Reuse orcSchema when push down Orc predicates
• [SPARK-39511] [SQL] Enhance push down local limit 1 for right side of left semi/anti join if join condition is empty
• [SPARK-38614] [SQL] Don’t push down limit through window that’s using percent_rank
• [SPARK-39551] [SQL] Add AQE invalid plan check
• [SPARK-39383] [SQL] Support DEFAULT columns in ALTER TABLE ADD COLUMNS to V2 data sources
• [SPARK-39538] [SQL] Avoid creating unnecessary SLF4J Logger
• [SPARK-39383] [SQL] Manual backport to DBR 11.x: Refactor DEFAULT column support to skip passing the primary Analyzer around
• [SPARK-39397] [SQL] Relax AliasAwareOutputExpression to support alias with expression
• [SPARK-39496] [SQL] Handle null struct in Inline.eval
• [SPARK-39545] [SQL] Override concat method for ExpressionSet in Scala 2.13 to improve the performance
• [SPARK-39340] [SQL] DS v2 agg pushdown should allow dots in the name of top-level columns
• [SPARK-39488] [SQL] Simplify the error handling of TempResolvedColumn
• [SPARK-38846] [SQL] Add explicit data mapping between Teradata Numeric Type and Spark DecimalType
• [SPARK-39520] [SQL] Override -- method for ExpressionSet in Scala 2.13
• [SPARK-39470] [SQL] Support cast of ANSI intervals to decimals
• [SPARK-39477] [SQL] Remove “Number of queries” info from the golden files of SQLQueryTestSuite
• [SPARK-39419] [SQL] Fix ArraySort to throw an exception when the comparator returns null
• [SPARK-39061] [SQL] Set nullable correctly for Inline output attributes
• [SPARK-39320] [SQL] Support aggregate function MEDIAN
• [SPARK-39261] [CORE] Improve newline formatting for error messages
• [SPARK-39355] [SQL] Single column uses quoted to construct UnresolvedAttribute
• [SPARK-39351] [SQL] SHOW CREATE TABLE should redact properties
• [SPARK-37623] [SQL] Support ANSI Aggregate Function: regr_intercept
• [SPARK-39374] [SQL] Improve error message for user specified column list
• [SPARK-39255] [SQL][3.3] Improve error messages
• [SPARK-39321] [SQL] Refactor TryCast to use RuntimeReplaceable
• [SPARK-39406] [PYTHON] Accept NumPy array in createDataFrame
• [SPARK-39267] [SQL] Clean up dsl unnecessary symbol
• [SPARK-39171] [SQL] Unify the Cast expression
• [SPARK-28330] [SQL] Support ANSI SQL: result offset clause in query expression
• [SPARK-39203] [SQL] Rewrite table location to absolute URI based on database URI
• [SPARK-39313] [SQL] toCatalystOrdering should fail if V2Expression can not be translated
• [SPARK-39301] [SQL][PYTHON] Leverage LocalRelation and respect Arrow batch size in createDataFrame with Arrow optimization
• [SPARK-39400] [SQL] spark-sql should remove hive resource dir in all case

Maintenance updates

See Databricks Runtime 11.2 maintenance updates.

System environment

• Operating System: Ubuntu 20.04.4 LTS
• Java: Zulu 8.56.0.21-CA-linux64
• Scala: 2.12.14
• Python: 3.9.5
• R: 4.1.3
• Delta Lake: 2.1.0

Installed Python libraries

Installed R libraries

R libraries are installed from the Microsoft CRAN snapshot on 2022-08-15.

Installed Java and Scala libraries (Scala 2.12 cluster version)