3.0.4.0

New Features

• Add SimpleDBConnectionPool to create and manage connection pools.
• Add support in TableAppender, TableUpserter, PartitionedTableAppender, and MultithreadedTableWriter for inserting data into stream tables in Orca.
• Add tryReconnectNums parameter to MultithreadedTableWriter to set reconnection attempts.
• Add usePublicName parameter to DBConnectionPool and MultithreadedTableWriter to support using node publicName as the connection address in a cluster environment.

Improvements

• Optimize log output.
• Optimize the reconnection mechanism to avoid holding the GIL in certain cases.
• Optimize Session connection information updates to timely detect node changes.
• Optimize script execution restrictions: when fetchSize is specified, other scripts cannot be executed until the current data fetch is completed.

Bug Fixes

• Fixed an issue where the cluster cannot be successfully connected when specifying tryReconnectNums=1.

3.0.3.0

Improvements

• Python 3.6 and 3.7 are no longer supported; Python 3.13 (with-GIL) is now supported.
• Added support for NumPy 2.x.
• Added new parameter enableStreamTableTimestamp for MultithreadedTableWriter to support writing data to stream tables with a timestamp column set via setStreamTableTimestamp.
• createTable has been renamed to createDimensionTable. The original name is retained as an alias for compatibility.
• Added a parser parameter to Session and DBConnectionPool to specify the script parser type used by the server after connecting.
• Improved logging for connection retries by promoting certain INFO messages to ERROR and setting the default log level to WARNING.

3.0.2.4

3.0.2.4

Improvements

• The filter parameter of subscribe now supports lambda expressions as strings.

Bug Fixes

• Fixed an issue in PROTOCOL_DDB where downloading a BOOL type column with more than 8,192 rows resulted in an error.

3.0.2.3

3.0.2.3

New Features

• Added get_types_from_schema for extracting table schema information. Additionally, __DolphinDB_Type__ now accepts the str type for column types.

Bug Fixes

• Fixed an issue where runTaskAsync would continuously make reconnection attempts instead of throwing an exception when executed by a guest user.

3.0.2.2

New Features

• Added support for Python 3.12.
• Introduced the Logger package, which allows logging information and print outputs to be directed to a specified location.
• The MultithreadedTableWriter now includes a reconnect parameter, which controls automatic reconnection if the network connection is interrupted during writing.

Improvements

• The resubTimeout parameter in subscribe has been renamed to resubscribeInterval.

Bug Fixes

• Decimal data upload behavior has been standardized to rounding.
• Fixed write performance degradation in the tableInsert and MultithreadedTableWriter, introduced in version 3.0.2.1.
• Fixed an issue where nodes in high-availability mode failed to switch properly.

3.0.2.1

New Features

• Added the io module with serialization and deserialization functions (dump, load, dumps, loads).

Improvements

• Enhanced error message clarity.

3.0.2.0

Improvements

• Enhanced parameter validation for methods such as Session.run.
• Added tryReconnectNums parameter to Session.connect method and DBConnectionPool constructor to set reconnection attempts.
• Added sqlStd parameter to Session and DBConnectionPool constructors to specify server-side SQL dialect.
• Changed default protocol in Session and DBConnectionPool constructors from PROTOCOL_PICKLE to PROTOCOL_DDB.
• When uploading a pandas.DataFrame, columns of string[pyarrow] type can now be specified as SYMBOL type.
• Added new parameters readTimeout and writeTimeout to Session.connect to specify API read and write timeouts.

3.0.1.1

Improvements

• The default parallelism parameter for the run method in both DBConnectionPool and Session classes has been increased from 2 to 64.

3.0.1.0

New Features

• New parameters primaryKey and indexes have been added to Database.createTable and Database.createPartitionedTable to support creating tables under the primary key engine.

Improvements

• Added support for verifying non-standard versions of third-party libraries (e.g., pandas 2.2.0rc0 ).
• Added support for automatic precision alignment when writing Decimal values.
• Optimized the performance of the append method in the Table class.
• Version dependency restriction: NumPy 2.0 is not currently supported.

Bug Fixes

• [CEP] Fixed a segmentation fault that occurred when EventClient subscribed to a table containing a time column not specified in the eventTimeFields parameter of the EventClient constructor.
• Fixed an issue that when writing data through tableUpsert, the write operation failed if a session reconnection occurred.

3.0.0.2

Improvements

• Added parameter disableDecimal for the run method of Session/session and DBConnectionPool classes, which determines whether to convert DECIMAL columns to DOUBLE type when downloading data.