v0.35.0

What's Changed

• Post release bump to 0.35.0-SNAPSHOT by @devinrsmith in #5424
• Update README_PyPi.md by @alexpeters1208 in #5428
• fix: Could not re-open application created variables by @mofojed in #5435
• Allow parallel TableLocation factory invocation in AbstractTableLocationProvider by @rcaudy in #5432
• Move Script Cache Clearing to Dagger by @nbauernfeind in #5436
• Update web version 0.75.0 by @github-actions in #5439
• Improve FileUtils.deleteRecursively diagnostics. by @cpwright in #5430
• Added S3 test tools to help with directory uploading by @devinrsmith in #5442
• Collect all pydeephaven-ticking wheels into single directory by @devinrsmith in #5444
• JS API should return nulls rather than undefined when reading table data by @niloc132 in #5437
• Fix SQL ArrayType / CustomType failure mode by @devinrsmith in #5445
• Update web version 0.75.1 by @github-actions in #5448
• Test for null result in DataIndexer.getOptimalPartialIndex by @rcaudy in #5449
• Update web version 0.76.0 by @github-actions in #5452
• Add logging when a ReplayTableBase has an error during refresh by @rcaudy in #5454
• Arrow Encoding for LocalDate and LocalTime by @nbauernfeind in #5446
• Fix inconsistencies in aggregations and query library functions. by @chipkent in #5368
• Make DataIndex accumulation and transformation parallelizable by @rcaudy in #5457
• Make Numeric.sum return null when there are no inputs or all null inputs by @chipkent in #5461
• Update web version 0.77.0 by @github-actions in #5463
• Fixed bug in cumsum and cumprod when there are multiple initial null values by @chipkent in #5464
• Widen returned types for UpdateBy floating point operations. by @lbooker42 in #5371
• Improved threading capabilities of S3+parquet by @malhotrashivam in #5451
• ClockFilter on an empty Table causes NPE on next update cycle by @cpwright in #5474
• Fix a numpy/seq incompatibility bug in 3.8 by @jmao-denver in #5466
• Provide Gradle toolchain resolver implementation by @niloc132 in #5475
• UPDATE: Fix formatting of a hyperlink in pydoc by @elijahpetty in #5356
• UPDATE: fix misspelling by @elijahpetty in #5354
• Bump base images by @devinrsmith in #5481
• Bump cla-assistant/github-action from 2.1.3.pre.beta to 2.4.0 by @dependabot in #5483
• Deprecate implicit String-to-UTF8-bytes ticket methods by @devinrsmith in #5405
• Fix Local Smoke Tests by Allowing ScriptSessionCacheInit's mkdir to Race in Parallel Tests by @nbauernfeind in #5465
• Added a soft reference based shared cache for S3 reads by @malhotrashivam in #5357
• Barrage Ingestion: Support TimeUnit Conversion on Instant and ZonedDateTime Columns by @nbauernfeind in #5488
• Publish deephaven-server-test jar by @devinrsmith in #5467
• Add a C++ client concurrent test/example by @jcferretti in #5490
• Cancel subscribeToLogs and messageStream RPCs when the Session closes by @devinrsmith in #5433
• Don't close long-lived RowSets in Downsampler BucketState by @niloc132 in #5478
• Implement DisplayWrapper#hashCode to fix SelectDistinct Behavior of Previewed Columns by @nbauernfeind in #5281
• Free performance data when ExportObject is released by @niloc132 in #5480
• Bugfix in URI conversion when working with directories by @malhotrashivam in #5493
• Update web version 0.78.0 by @github-actions in #5500
• Move away from ClientMiddleware and ClientAuthHandler in pydeephaven by @jcferretti in #5489
• Unit tests for S3 + parquet by @malhotrashivam in #5441
• Barrage Ingestion: Fix ClassCastExceptions on Empty Primitive Arrays by @nbauernfeind in #5503
• Small ring chunk source cleanup by @devinrsmith in #5502
• Fixed the parquet uint32 reading failure by @malhotrashivam in #5495
• Force PartitionAwareSourceTable to coalesce whenever a partition filter is applied via where by @rcaudy in #5485
• Migrate properties to netty-app from jetty-app and update proj status by @niloc132 in #5505
• 5414 reconnect subscriptions by @niloc132 in #5501
• FlightMessageRoundTripTest: Inner List Type Should not Have Deephaven Type Metadata by @nbauernfeind in #5504
• Generate grpc/pb with commonjs_strict by @niloc132 in #5336
• Use concurrent queue for InputTableService by @devinrsmith in #5492
• Factory for Vectors over entire columns from a Table by @rcaudy in #5506
• Add java-client support for primitive array types by @devinrsmith in #5494
• Updated guava dependency version from 32.0.1 to 33.2.0 by @malhotrashivam in #5511
• Add All-AI deployments by @jjbrosnan in #5367
• Added support for reading Parquet V2 pages by @malhotrashivam in #5508
• Add release note about patch release quick check by @devinrsmith in #5507
• Added fix for NPE when reading from S3 by @malhotrashivam in #5514
• Fix for excessive memory usage when reading parquet by @malhotrashivam in #5516
• Upgraded version for parquet-hadoop by @malhotrashivam in #5469
• Mitigate risks of SoftReference cache OOM by @devinrsmith in #5520
• Bump base server image, pick up jpy==0.17.0 by @devinrsmith in #5527
• Update web version 0.79.0 by @github-actions in #5531
• Remove branching in UDF wrappers via. code gen by @jmao-denver in #5487
• Update C++ dependencies, roughly matching ones used in python. by @jcferretti in #5472
• fix: PyPi README should indicate we require 3.8 or newer by @mofojed in #5544
• Barrage: Coalesce UncoalescedTables Prior to Subscribing/Fetching Data by @nbauernfeind in #5541
• Cache size of files fetched from S3 by @malhotrashivam in #5545
• Proposed superficial fixes around MutableInt/MutableLong usage by @niloc132 in #5450
• Temporary fix for broken US cran mirror for R packages by @niloc132 in #5549
• Ensure that ConsoleService.getHeapInfo is authenticated by @devinrsmith in #5536
• Fix publication details for shadowed jars by @niloc132 in #5548
• fix: Expose layoutHints in JS API for TreeTable by @mofojed in #5555
• Update web version 0.80.0 by @github-actions in #5556
• StorageService paths should be treated as opaque strings, part 1 by @niloc132 in #5532
• Generate ToChar/Short/ByteFunction interfaces and use in chunk apis by @niloc132 in #5559
• fix!: wavg() and wsum() Numeric functions return null when provided a vector containing only null values by @lbooker42 in #5524
• Fix names of type annotations not matching between definition and use. by @jcferretti in #5562
• Update web version 0.81.1 by @github-actions in #5570
• Address initial lastNotificationStep issue for TimeTable and ReplayTableBase. Also correct TimeTable satisfaction testing. by @rcaudy in #5577
• Initial commit of Iceberg integration. by @lbooker42 in #5277
• Fix SourceTable coalesce bug when indexed columns are dropped by @rcaudy in #5578
• Retire DataColumn and replace usages by @rcaudy in #5553
• [Breaking] Deleted deprecated parquet read/write APIs by @malhotrashivam in #5510
• Upgrade aircompressor from v0.25 to v0.27 by @malhotrashivam in #5568
• Cleanup to BarrageStreamGenerator and its related classes in antipation of porting to JS by @niloc132 in #5552
• Remove dtypes.int_/float_/int_array/float_array to eliminate confusion by @jmao-denver in #5580
• BaseArrayBackedInputTable user safety fixes by @rcaudy in #5588
• Address missing DateTimeUtils functionality for LocalDateTime by @rcaudy in #5540
• Update to GWT 2.11, clean up old gradle wiring by @niloc132 in #5576
• table_definition will be optional for parquet batch_write by @malhotrashivam in #5586
• Adds docker registry projects for localstack and minio by @devinrsmith in #5593
• Fix an issue with boxed type arg matching and disable np scalar warnings for numba by @jmao-denver in #5599
• Introduce runtime version checks for flatbuffer compatibility by @niloc132 in #5583
• Fix csv.py parser mappings by @devinrsmith in #5602
• Make region optional for S3 instructions by @malhotrashivam in #5603
• fix: Modify getKeyColumns method for JsInputTable by @AkshatJawne in #5600
• CSV writer improvements by @rcaudy in #5604
• fix: Correctly pad zeroes for subsecond datetime formats by @niloc132 in #5612
• fix: JS InputTable test failure and race by @niloc132 in #5616
• Add instructions for patching old releases by @stanbrub in #5597
• Improvements to buffered reading for parquet by @malhotrashivam in #5611
• Update web version 0.82.0 by @github-actions in #5598
• Fix some issues in the test code found by numpy 2.0.0 by @jmao-denver in #5621
• Add a new init parameter to the embedded server to allow the user to only override the Deephaven recommended JVM args by @jmao-denver in #5542
• ci: Trigger actions on pull requests to rc/v* branches by @mofojed in #5626
• Optimizations in parquet file page materialization by @malhotrashivam in #5582
• Bump server base images for numpy 2 by @devinrsmith in #5623
• JSON configuration and Jackson Streaming Object Processor by @devinrsmith in #5225
• Add defensive bearer token rotation for java client by @devinrsmith in #5539
• feat: Added query library functions to overload Period and Duration arithmetic by @chipkent in #5509
• fix: Remove table attribute size cutoff limit by @mofojed in #5628
• fix: signal.pause is not available on Windows by @mofojed in #5618
• Add a listener that only fires when all dependencies are satisfied by @jmao-denver in #5571
• fix!: statistical functions should return null when provided a vector of only null values by @lbooker42 in #5606
• Optimizations for BigDecimal and BigInteger parquet reads by @malhotrashivam in #5638
• Allow extra headers in Barrage session creation by @arman-ddl in #5637
• ErrorTransformer: prevent NPE by @nbauernfeind in #5633
• [BREAKING] Fix for OOM error when reading from S3 by @malhotrashivam in #5613
• Validate arg type for sort/partition_by by @jmao-denver in #5652
• Fix getNextRowSequenceWithLength callsites that close by @devinrsmith in #5653
• feat!: Remove deprecated Daylight Savings Time (DST) methods by @chipkent in #5266
• feat: expose Iceberg features to python users by @lbooker42 in #5590
• Change py-embedded-server from WARN to INFO logging by @devinrsmith in #5644
• fix: correct weighted summation null handling behavior by @lbooker42 in #5660
• fix: correct OOB test failures from mismatching Iceberg API table column names by @lbooker42 in #5662
• Remove server dependency on extensions-iceberg by @devinrsmith in #5664
• fix: rename iceberg:s3 sub-project to iceberg-s3. by @lbooker42 in #5663
• Fix ExactMatch Filter for Non-Convertible Types; Use RangeFilter on QueryScope Vars by @nbauernfeind in #5587
• fix: Only update the redir index on the first column by @niloc132 in #5659
• Update web version 0.83.0 by @github-actions in #5667
• fix: Modify key table viewport to only include key columns by @AkshatJawne in #5669
• feat: reduce included dependencies for Iceberg implementation by @lbooker42 in #5668
• Further improvements to parquet page materializaition by @malhotrashivam in #5670
• C++ Client: End of Barrage stream should throw; reformat exception messages by @kosak in #5682
• C++ Client: fix WAvg, fix noexcept, optimize PercentileBy, fix comments by @kosak in #5684
• Instant Column gRPC Filters: Fix Range Quick Filters and use MatchFilter over ConditionFilter by @nbauernfeind in #5671

New Contributors

• @AkshatJawne made their first contribution in #5600

Full Changelog: v0.34.3...v0.35.0