ReleaseBytes
databricks Databricks Python SDK Releases · · 0.134.0

Databricks Python SDK v0.134.0 Released with AI Functions and Pipeline Updates

sdkaidatabricksengineermedia
breaking patch

Databricks Python SDK v0.134.0 is now available, introducing significant updates including new `aifunctions` and `sandbox` packages. A breaking change makes `source_table` and `source_schema` fields optional in the Pipelines API, enhancing flexibility for diverse data sources like streaming connectors. The release also adds numerous API methods and fields across services like IAM, feature engineering, and Postgres snapshot management, improving developer capabilities. Developers should review code relying on the previously required pipeline fields to handle their potential absence.

  • →Change source_schema field for databricks.sdk.service.pipelines.SchemaSpec to no longer be required
  • →Change source_table field for databricks.sdk.service.pipelines.TableSpec to no longer be required
  • →Remove disabled field for databricks.sdk.service.catalog.InferenceTableConfig
  • →Remove owner field for databricks.sdk.service.catalog.McpService
  • →Remove owner field for databricks.sdk.service.catalog.ModelProviderService
Breaking changes (15) ›

  • TableSpec.source_table and SchemaSpec.source_schema in the Pipelines API (databricks.sdk.service.pipelines) are now optional. Previously required, they may be absent for source types — such as streaming or message-bus connectors — that do not use a source table or schema name. Code that assumes these fields are always present should handle their absence

  • Change source_schema field for databricks.sdk.service.pipelines.SchemaSpec to no longer be required
  • Change source_table field for databricks.sdk.service.pipelines.TableSpec to no longer be required
  • Remove disabled field for databricks.sdk.service.catalog.InferenceTableConfig
  • Remove owner field for databricks.sdk.service.catalog.McpService
  • Remove owner field for databricks.sdk.service.catalog.ModelProviderService
  • Remove plan_type field for databricks.sdk.service.catalog.ModelProviderServiceConfigAnthropicProviderRelayedConfig
  • Remove owner field for databricks.sdk.service.catalog.ModelService
  • Remove traffic_splitting field for databricks.sdk.service.catalog.ModelServiceConfigRoutingConfig
  • Remove entity_columns and timeseries_column fields for databricks.sdk.service.ml.DeltaTableSource
  • Remove filter_condition, inputs and time_window fields for databricks.sdk.service.ml.Feature
  • Remove extra_parameters and function_type fields for databricks.sdk.service.ml.Function
  • Remove entity_column_identifiers and timeseries_column_identifier fields for databricks.sdk.service.ml.KafkaSource
  • Remove cron_schedule field for databricks.sdk.service.ml.MaterializedFeature
  • Remove continuous field for databricks.sdk.service.ml.TimeWindow
Enhancements (35) ›
  • Add databricks.sdk.service.aifunctions and databricks.sdk.service.sandbox packages
  • Add w.ai_functions workspace-level service
  • Add w.sandbox workspace-level service
  • Add update_deployment() method for w.bundle_deployments workspace-level service
  • Add genie_cancel_response() method for w.genie workspace-level service
  • Add get_external_group(), get_external_service_principal() and get_external_user() methods for a.account_iam_v2 account-level service

  • Add get_external_group_proxy(), get_external_service_principal_proxy() and get_external_user_proxy() methods for w.workspace_iam_v2 workspace-level service

  • Add backfill_features(), cancel_operation() and get_operation() methods for w.feature_engineering workspace-level service

  • Add create_snapshot(), delete_snapshot(), get_snapshot(), get_snapshot_schedule(), list_snapshots() and update_snapshot_schedule() methods for w.postgres workspace-level service

  • Add assume_group_id field for databricks.sdk.service.apps.Space
  • Add operations field for databricks.sdk.service.bundledeployments.Version
  • Add function_arg_expression field for databricks.sdk.service.catalog.FunctionArgument
  • Add aws_context_id field for databricks.sdk.service.compute.NodeTypeFlexibility
  • Add parameters field for databricks.sdk.service.jobs.AlertTask
  • Add provisioned_capacity_id field for databricks.sdk.service.jobs.ComputeSpec
  • Add maintenance_window field for databricks.sdk.service.jobs.Continuous
  • Add maintenance_window field for databricks.sdk.service.jobs.ContinuousTriggerConfiguration
  • Add lateness field for databricks.sdk.service.ml.DataSource
  • Add latest_backfill_operation field for databricks.sdk.service.ml.MaterializedFeature
  • Add delay and offset fields for databricks.sdk.service.ml.SlidingWindow
  • Add excluded_columns and record_type_filter fields for databricks.sdk.service.ml.Stream
  • Add start_time field for databricks.sdk.service.ml.TimeWindow
  • Add delay and offset fields for databricks.sdk.service.ml.TumblingWindow
  • Add rabbitmq_options field for databricks.sdk.service.pipelines.ConnectorOptions
  • Add source_snapshot field for databricks.sdk.service.postgres.BranchSpec
  • Add source_snapshot field for databricks.sdk.service.postgres.BranchStatus

  • Add deployment_resource_type_vector_search_endpoint, deployment_resource_type_vector_search_index, deployment_resource_type_job_run, deployment_resource_type_postgres_catalog, deployment_resource_type_postgres_synced_table, deployment_resource_type_genie_space, deployment_resource_type_instance_pool, deployment_resource_type_postgres_database, deployment_resource_type_postgres_role, deployment_resource_type_secret and deployment_resource_type_cluster_policy enum values for databricks.sdk.service.bundledeployments.DeploymentResourceType

  • Add operation_status_pending enum value for databricks.sdk.service.bundledeployments.OperationStatus
  • Add aws_secrets_manager and azure_key_vault enum values for databricks.sdk.service.catalog.ConnectionType
  • Add external_use_location enum value for databricks.sdk.service.catalog.Privilege
  • Add model, model_service, mcp_service and model_provider_service enum values for databricks.sdk.service.catalog.SecurableType
  • Add gpu_1x_h100 and gpu_8x_b300 enum values for databricks.sdk.service.compute.HardwareAcceleratorType
  • Add rabbitmq enum value for databricks.sdk.service.pipelines.IngestionSourceType
  • Change source_schema field for databricks.sdk.service.pipelines.SchemaSpec to no longer be required
  • Change source_table field for databricks.sdk.service.pipelines.TableSpec to no longer be required
Read the original announcement →

https://github.com/databricks/databricks-sdk-py/releases/tag/v0.134.0

Related releases