• [BREAKING-EXPERIMENTAL] The batch_data attribute of BatchRequest has been removed. To pass in in-memory dataframes at runtime, the new RuntimeDataConnector should be used
• [BREAKING-EXPERIMENTAL] RuntimeDataConnector must now be passed Batch Requests of type RuntimeBatchRequest
• [BREAKING-EXPERIMENTAL] The PartitionDefinitionSubset class has been removed - the parent class IDDict is used in its place
• [BREAKING-EXPERIMENTAL] partition_request was renamed data_connector_query. The related PartitionRequest class has been removed - the parent class IDDict is used in its place
• [BREAKING-EXPERIMENTAL] partition_definition was renamed batch_identifiers. The related PartitionDefinition class has been removed - the parent class IDDict is used in its place
• [BREAKING-EXPERIMENTAL] The PartitionQuery class has been renamed to BatchFilter
• [BREAKING-EXPERIMENTAL] The batch_identifiers key on DataConnectorQuery (formerly PartitionRequest) has been changed to batch_filter_parameters
• [ENHANCEMENT] Added a new RuntimeBatchRequest class, which can be used alongside RuntimeDataConnector to specify batches at runtime with either an in-memory dataframe, path (filesystem or s3), or sql query
• [ENHANCEMENT] Added a new RuntimeQueryBatchSpec class
• [BUGFIX] Fixed issue where Sorters were not being applied correctly when data_connector_query contained limit or index #2617
• [DOCS] Updated docs to reflect above class name changes
• [DOCS] Added the following docs: "How to configure sorting in Data Connectors", "How to configure a Runtime Data Connector", "How to create a Batch Request using an Active Data Connector", "How to load a database table, view, or query result as a Batch"
• [DOCS] Updated the V3 API section of the following docs: "How to load a Pandas DataFrame as a Batch", "How to load a Spark DataFrame as a Batch"