- Release: #1022 fixes for pandas 2 and latest numpy
- Compatibility: unpin pandas and numpy
- Compatibility: update pinning to pymongo<4
- Compatibility: tested on python 3.11
- Release: #1022 Prepare release 1.82.2
- Perf: #1021 Improve performance of Tickstore read
- Bugfix: #1019 Remove failing requirements.txt versions
- Docs: #1019 Updated links to ArcticDB website and docs
- Docs: #1017 Fix doc for chunkstore's list_symbols
- Release: #1014 Prepare release 1.82.1
- Bugfix: #1013 Fix serialized tz name for index timezones of dateutil.timezone.tzutc()
- Bugfix: #1013 Fix serialized tz name for single-level index timezones of dateutil.timezone.gettz("UTC")
- Release: #1012 version 1.82.0
- Compatibility: update pinning from pandas<1.1, to pandas<2
- Compatibility: update pinning from numpy<1.19 to numpy<2
- Build: specify which version of python/numpy/pandas to test
- Bugfix: #887 #928 resolve import error on pandas.Panel, Panel was removed in pandas 1.0
- Bugfix: fix pandas get_timezone() argument cannot be None
- Bugfix: fix np.float/np.int/np.book depreciations
- Bugfix: fix numpy tostring() is deprecated. Use tobytes() instead.
- Bugfix: numpy.testing.utils is deprecated
- Bugfix: pandas.util.testing is deprecated
- Cleanup: remove async_benchmark
- Feature: #1000, #1001, #1002, #1003, #1007, #1008 ArcticDB docs, links etc
- Bugfix: #1009 force CircleCI to install tzlocal==4.2
- Feature: #996 release wheel to pypi as well as tar.gz, egg.
- Bugfix: #979 fix circleci mongo install
- Feature: #980 get testpack running with mongod 4.4.18
- Bugfix: #985 suppress FutureWarning for Panel import
- Bugfix: #989 VersionStore slow snapshot delete add new index
- Bugfix: #806, #990 DateTimeIndex tz serialization
- Feature: #950 remove all traces of python 2.7 and six package
- Feature: #959 add python 3.8 support
- Feature: #962 pin pandas<1.1.0, numpy<1.19.0
- Bugfix: #940 fix rows per chunk causing divide by zero
- Feature: #943 check markdown rendering of README.md and CHANGES.md in build
- Bugfix: #946 move test dependencies to tests_require
- Feature: #941 use named index for VersionStore to avoid 127 max fully qualified index name
- Bugfix: #935, #936 fixed README.md rendering which was preventing pypi upload
- Bugfix: #932 revert serialization-optimization (#909, #910)
- Bugfix: #855 use IXSCAN for list_symbols which speeds up snapshotting (actually #856)
- Bugfix: #926 avoid pathologically slow count_documents() call with pymongo > 3.6.0
- Feature: #919 Add CircleCI badge to README.md
- Feature: #917 Add CircleCI build
- Bugfix: #910 Fix column subsetting bug
- Bugfix: #909 Speedup FrameToArray serialiser for ChunkStore by removing intermediate DataFrame construction
- Bugfix: #902 Chunkstore-read-speedup
- Bugfix: #872 Do not spam if not permissioned on cache db
- Bugfix: #874 Pickle protocol 5 not supported in 3.7 and below
- Bugfix: Handle uninitialized cache object
- Feature: #897 Pin numpy 1.18.4
- Feature: #897 Pin pandas 1.0.3
- Bugfix: #881 Fix for issue #815
- Bugfix: #875 Fix flake8 errors
- Bugfix: Update code to work with latest version of pandas
- Bugfix: Fix missing column handling in numpy serializer
- Bugfix: #832 Bytes to unicode for pd.Series
- Feature: #825 Max BSON size configurable
- Bugfix: #823 Don't retry bulk write errors
- Feature: #802 kwargs passed through to iterators in chunkstore
- Feature: Allow list_library caching to be tunable via a settings collection.
- Bugfix: #777: Handle empty columns in dataframes
- Bugfix: Pandas convert multiindexes to unicode if flag set.
- Bugfix: #765 Pandas data columns require encoding to convert to unicode.
- Bugfix: Use the default encoding (ascii) instead of bytes for unpickling.
- Bugfix: pypandoc not rendering README correctly for PYPI
- Bugfix: #744 get_info on an empty dataframe raises an exception
- Bugfix: #751 Handle unpickling py2 datetime and other object types with latin1
- Bugfix: #598 Add an option to convert py2 str to py3 str (unicode type) on read.
- Bugfix: #712 Pandas deprecation warning in chunkstore serializer
- BugFix: #670 Lots of pycodestyle fixes
- BugFix: #691 Fix arrays_to_mgr for pandas 0.23.4+
- BugFix: #698 Fix no handler in logging warning
- BugFix: #705 Use sort_index instead of sortlevel_
- BugFix: #710: Initialize SEGMENT_SHA in versions for writes and appends
- Feature: #669 Experimental initial implementation of async arctic
- Feature: #704 Do not enable sharding by default in BSONStore.
- Feature: #688 Remove Chunkstore check for issue #422
- Feature: #717 Add a best effort cache for list_libraries
- Bugfix: #722 Exception while reading a column from chunkstore with a column missing in the chunk
- Bugfix: #658 Write/append errors for Panel objects from older pandas versions
- Feature: #653 Add version meta-info in arctic module
- Feature: #663 Include arctic numerical version in the metadata of the version document
- Feature: #650 Implemented forward pointers for chunks in VersionStore (modes: enabled/disabled/hybrid)
- Feature: #577 Added implementation for incremental serializer for numpy records
- Bugfix: #648 Fix issue with Timezone aware Pandas types, which don't contain hasobject attribute
- Bugfix: #645 Fix write errors for Pandas DataFrame that has mixed object/string types in multi-index column
- Bugfix: #157 Assure that serialized dataframes remain value-equivalent (e.g. avoid NaN --> 'nan' in mixed string columns)
- Bugfix: #608 Ensure Arctic performs well with MongoDB 3.6 (sorting)
- Bugfix: #629 Column kwarg no longer modified
- Bugfix: #641 DateRange.intersection open/closed range fix
- Feature: #493 Can pass kwargs when calling MongoClient, e.g. for ssl
- Feature: #590 Faster write handler selection for DataFrames with objects
- Feature: #604 Improved handling handling for pickling serialization decidions
- Docs: VersionStore documentation
- Bugfix: Issue #612 ThreadPool should be created by process using it
- Feature: Upsert option on appends in ChunkStore
- Feature: #553 Compatibility with both the new and old LZ4 API
- Feature: #571 Removed the Cython LZ4 code, use the latest python-lz4
- Feature: #557 Threadpool based compression. Speed imrpovement and tuning benchmarks.
- Bugfix: fix tickstore unicode handling, support both unicode and utf-8 arrays
- Bugfix: #591 Fix tickstore reads not returning index with localized timezone
- Feature: #595 add host attribute to VersionedItem.
- Bugfix: #594 Enable sharding on chunkstore
- Bugfix: #579 Fix symbol corruption due to restore_version and append
- Bugfix: #584 Fix list_versions for a snapshot after deleting symbols in later versions
- Bugfix: #561 Fix PickleStore read corruption after write_metadata
- Bugfix: #168 Do not allow empty string as a column name
- Bugfix: #483 Remove potential floating point error from datetime_to_ms
- Bugfix: #271 Log when library doesnt exist on delete
- Feature: MetaDataStore: added list_symbols with regexp, as_of and metadata fields matching filters
- Feature: Support for serialization of DataFrames in Pandas 0.23.x
- Bugfix: #534 VersionStore: overwriting a symbol with different dtype (but same data format) does not raise exceptions anymore
- Bugfix: #531 arctic_prune_versions: clean broken snapshot references before pruning
- Bugfix: setup.py develop in a conda environment on Mac
- Feature: #490 add support to numpy 1.14
- Bugfix: #521 Clang 6.0 compiler support on macOS
- Feature: #510 VersionStore: support multi column in pandas DataFrames
- Bugfix: #517 VersionStore: append does not duplicate data in certain corner cases
- Bugfix: #519 VersionStore: list_symbols speed improvement and fix for memory limit exceed
- Feature: #288 Mapping reads and writes over chunks in chunkstore
- Bugfix: #508 VersionStore: list_symbols and read now always returns latest version
- Bugfix: #512 Improved performance for list_versions
- Bugfix: #515 VersionStore: _prune_previous_versions now retries the cleanup operation
- Bugfix: #503 ChunkStore: speedup check for -1 segments
- Feature: #504 Increasing number of libraries in Arctic to 5000.
- Bugfix: Increase performance of invalid segment check in chunkstore
- Bugfix: #501 Fix the spurious data integrity exceptions at write path, due to moving chunks form the balancer
- Bugfix: #491 roll back the use of frombuffer to fromstring, fixes the read-only ndarray issue
- Feature: #206 String support for tickstore
- Bugfix: #486 improve mongo_retry robustness with failures for version store write/append
- Bugfix: #468 Re-adding compatibility for pandas 0.20.x
- Bugfix: #476 Ensure we re-auth when a new MongoClient is created after fork
- Bugfix: #439 fix cursor timeouts in chunkstore iterator
- Bugfix: #450 fix error in chunkstore delete when chunk range produces empty df
- Bugfix: #442 fix incorrect segment values in multi segment chunks in chunkstore
- Feature: #457 enchances fix for #442 via segment_id_repair tool
- Bugfix: #385 exceptions during quota statistics no longer kill a write
- Feature: PR#161 TickStore.max_date now returns a datetime in the 'local' timezone
- Feature: #425 user defined metadata for tickstore
- Feature: #464 performance improvement by avoiding unnecessary re-authentication
- Bugfix: #250 Added multiprocessing safety, check for initialized MongoClient after fork.
- Feature: #465 Added fast operations for write only metadata and restore symbol to a version
- Bugfix: #440 Fix read empty MultiIndex+tz Series
- Perf: #408 Improve memory performance of version store's serializer
- Bugfix #394 Multi symbol read in chunkstore
- Bugfix: #407 Fix segment issue on appends in chunkstore
- Bugfix: Inconsistent returns on MetadataStore.append
- Bugfix: #412 pandas deprecation and #289 improve exception report in numpy record serializer
- Bugfix: #420 chunkstore ignoring open interval date ranges
- Bugfix: #427 chunkstore metadata not being correctly replaced during symbol overwrite
- Bugfix: #431 chunkstore iterators do not handle multi segment chunks correctly
- Bugfix: #397 Remove calls to deprecated methods in pymongo
- Bugfix: #402 Append to empty DF fails in VersionStore
- Feature: #396 MetadataStore.read now supports as_of argument
- Bugfix: #397 Pin pymongo==3.4.0
- Feature: #392 MetadataStore
- Bugfix: #384 sentinels missing time data on chunk start/ends in ChunkStore
- Bugfix: #382 Remove dependency on cython being pre-installed
- Bugfix: #343 Renaming libraries/collections within a namespace/database
- BugFix: Rollback #363, as it breaks multi-index dataframe
- Bugfix: #372 OSX build improvements
- Feature: Re-introduce #363
concat
flag, essentially undo-ing 1.45 - BugFix: #377 Fix broken
replace_one
on BSONStore and addbulk_write
- Feature: #374 Shard BSONStore on
_id
rather thansymbol
- BugFix: Rollback #363, which can cause ordering issues on append
- Feature: #364 Expose compressHC from internal arctic LZ4 and remove external LZ4 dependency
- Feature: #363 Appending older data (compare to what's exist in library) will raise. Use
concat=True
to append only the new bits - Feature: #371 Expose more functionality in BSONStore
- Bugfix: #350 remove deprecated pandas calls
- Bugfix: #360 version incorrect in empty append in VersionStore
- Feature: #365 add generic BSON store
- Bugfix: #346 fixed daterange subsetting error on very large dateframes in version store
- Bugfix: #351 $size queries can't use indexes, use alternative queries
- Bugfix: #334 Chunk range param with pandas object fails in chunkstore.get_chunk_ranges
- Bugfix: #339 Depending on lz4<=0.8.2 to fix build errors
- Bugfix: #342 fixed compilation errors on Mac OSX
- Bugfix: #344 fixed data corruption problem with concurrent appends
- BugFix: #330 Make Arctic._lock reentrant
- Feature: #329 Add reset() method to Arctic
- Bugfix: #324 Datetime indexes must be sorted in chunkstore
- Feature: #290 improve performance of tickstore column reads
- Bugfix: #300 to_datetime deprecated in pandas, use to_pydatetime instead
- Bugfix: #309 formatting change for DateRange
__str__
- Feature: #313 set and read user specified metadata in chunkstore
- Feature: #319 Audit log support in ChunkStor
- Bugfix: #216 Tickstore write fails with named index column
- Feature: Default to hashed based sharding
- Bugfix: retry socket errors during VersionStore snapshot operations
- Bugfix: #296 Cannot compress/decompress empty string
- Feature: #294 Move per-chunk metadata for chunkstore to a separate collection
- Bugfix: #292 Account for metadata size during size chunking in ChunkStore
- Feature: #283 Support for all pandas frequency strings in ChunkStore DateChunker
- Feature: #286 Add has_symbol to ChunkStore and support for partial symbol matching in list_symbols
- Feature: #275 Tuple range object support in DateChunker
- Bugfix: #273 Duplicate columns breaking serializer
- Feature: #267 Tickstore.delete returns deleted data
- Dependency: #266 Remove pytest-dbfixtures in favor of pytest-server-fixtures
- Feature: #260 quota support on Chunkstore
- Bugfix: #259 prevent write of unnamed columns/indexes
- Bugfix: #252 pandas 0.19.0 compatibility fixes
- Bugfix: #249 open ended range reads on data without index fail
- Bugfix: #262 VersionStore.append must check data is written correctly during repack
- Bugfix: #263 Quota: Improve the error message when near soft-quota limit
- Perf: #265 VersionStore.write / append don't aggressively add indexes on each write
- Bugfix: #247 segmentation read fix in chunkstore
- Feature: #243 add get_library_type method
- Bugfix: more cython changes to handle LZ4 errors properly
- Feature: #239 improve chunkstore's get_info method
- Feature: #235 method to return chunk ranges on a symbol in ChunkStore
- Feature: #234 Iterator access to ChunkStore
- Bugfix: #236 Cython not handling errors from LZ4 function calls
- Bugfix: #228 Mongo fail-over during append can leave a Version in an inconsistent state
- Feature: #193 Support for different Chunkers and Serializers by symbol in ChunkStore
- Feature: #220 Raise exception if older version of arctic attempts to read unsupported pickled data
- Feature: #219 and #220 Support for pickling large data (>2GB)
- Feature: #204 Add support for library renaming
- Feature: #209 Upsert capability in ChunkStore's update method
- Feature: #207 Support DatetimeIndexes in DateRange chunker
- Bugfix: #232 Don't raise during VersionStore #append(...) if the previous append failed
- Bugfix: #195 Top level tickstore write with list of dicts now works with timezone aware datetimes
- Bugfix: #187 Compatibility with latest version of pytest-dbfixtures
- Feature: #182 Improve ChunkStore read/write performance
- Feature: #162 Rename API for ChunkStore
- Feature: #186 chunk_range on update
- Bugfix: #189 range delete does not update symbol metadata
- Bugfix: Faster TickStore querying for multiple symbols simultaneously
- Bugfix: TickStore.read now respects
allow_secondary=True
- Bugfix: #147 Add get_info method to ChunkStore
- Bugfix: Periodically re-cache the library.quota to pick up any changes
- Bugfix: #166 Add index on SHA for ChunkStore
- Bugfix: #169 Dtype mismatch in chunkstore updates
- Feature: #171 allow deleting of values within a date range in ChunkStore
- Bugfix: #172 Fix date range bug when querying dates in the middle of chunks
- Bugfix: #176 Fix overwrite failures in Chunkstore
- Bugfix: #178 - Change how start/end dates are populated in the DB, also fix append so it works as expected.
- Bugfix: #43 - Remove dependency on hardcoded Linux timezone files
- Bugfix: Ensure that Tickstore.write doesn't allow out of order messages
- Bugfix: VersionStore.write now allows writing 'None' as a value
- Bugfix: Backwards compatibility reading/writing documents with previous versions of Arctic
- Bugfix: #109 Ensure stable sort during Arctic read
- Feature: New benchmark suite using ASV
- Bugfix: #129 Fixed an issue where some chunks could get skipped during a multiple-symbol TickStore read
- Bugfix: #135 Fix issue with different datatype returned from pymongo in python3
- Feature: #130 New Chunkstore storage type
- Bugfix: #106 Fix Pandas Panel storage for panels with different dimensions
- Feature: #98 Add initial_image as optional parameter on tickstore write()
- Bugfix: #100 Write error on end field when writing with pandas dataframes
- Feature: Add python 3.3/3.4 support
- Bugfix: #95 Fix raising NoDataFoundException across multiple low level libraries
- Bugfix: #81 Fix broken read of multi-index DataFrame written by old version of Arctic
- Bugfix: #49 Fix strifying tickstore
- Feature: Add timezone suppport to store multi-index dataframes
- Bugfix: Fixed broken sdist releases
- Feature: ArticTransaction now supports non-audited 'transactions':
audit=False
with ArcticTransaction(Arctic('hostname')['some_library'], 'symbol', audit=False) as at:
...
This is useful for batch jobs which read-modify-write and don't want to clash with
concurrent writers, and which don't require keeping all versions of a symbol.
- Feature: get_info API added to version_store.
- Bugfix: correct version detection for Pandas >= 0.18.
- Bugfix: retrying connection initialisation in case of an AutoReconnect failure.
- Bugfix: Improve performance of saving multi-index Pandas DataFrames by 9x
- Bugfix: authenticate should propagate non-OperationFailure exceptions (e.g. ConnectionFailure) as this might be indicative of socket failures
- Bugfix: return 'deleted' state in VersionStore.list_versions() so that callers can pick up on the head version being the delete-sentinel.
- Bugfix: VersionStore.read(date_range=...) could do the wrong thing with TimeZones (which aren't yet supported for date_range slicing.).
- Bugfix: fix authentication race condition when sharing an Arctic instance between multiple threads.
- Bugfix: compatibility with both 3.0 and pre-3.0 MongoDB for querying current authentications
- Feature: Add support for reading a subset of a pandas DataFrame in VersionStore.read by passing in an arctic.date.DateRange
- Bugfix: Reauth against admin if not auth'd against a library a specific library's DB. Sometimes we appear to miss admin DB auths. This is to workaround that until we work out what the issue is.
- Feature: Add support for multi-index Bitemporal DataFrame storage. This allows persisting data and changes within the DataFrame making it easier to see how old data has been revised over time.
- Bugfix: Ensure we call the error logging hook when exceptions occur
- Always use the primary cluster node for 'has_symbol()', it's safer
- Bugfixes for timezone handling, now ensures use of non-naive datetimes
- Bugfix for tickstore read missing images
- Improvements to command-line control scripts for users and libraries
- Bugfix for pickling top-level Arctic object
- Allow snapshotting a range of versions in the VersionStore, and snapshot all versions by default.
- Bugfix for backwards-compatible unpickling of bson-encoded data
- Added switch for enabling parallel lz4 compression
- Initial public release