ReleaseBytes
databricks Databricks Python SDK Releases · · 0.135.0

Databricks Python SDK v0.135.0 Introduces API Changes, Including Breaking Ones

sdkmldatabricksengineer
breaking patch

Databricks has released version 0.135.0 of its Python SDK, introducing several API changes, including both new functionality and breaking modifications. The update adds a synchronous command execution method for workspace sandbox services and new fields for ML model scheduling, feature definitions, and publishing specifications. Conversely, it removes several fields related to model service configuration routing and rate limiting from the Unity Catalog service. Developers using the SDK, particularly those interacting with ML or Unity Catalog services, will need to review these changes and potentially update their code.

  • →Remove first_token_timeout field for databricks.sdk.service.catalog.ModelServiceConfigRoutingConfig
  • →Remove request_tag_key and request_tag_value fields for databricks.sdk.service.catalog.RateLimit
  • →Remove rate_limit_key_request_tag enum value for databricks.sdk.service.catalog.RateLimitRateLimitKey
  • →Add execute_command_sync() method for w.sandbox workspace-level service
  • →Add mode field for databricks.sdk.service.ml.CronSchedule
Breaking changes (3) ›
  • Remove first_token_timeout field for databricks.sdk.service.catalog.ModelServiceConfigRoutingConfig
  • Remove request_tag_key and request_tag_value fields for databricks.sdk.service.catalog.RateLimit
  • Remove rate_limit_key_request_tag enum value for databricks.sdk.service.catalog.RateLimitRateLimitKey
Enhancements (4) ›
  • Add execute_command_sync() method for w.sandbox workspace-level service
  • Add mode field for databricks.sdk.service.ml.CronSchedule
  • Add time_window field for databricks.sdk.service.ml.Feature
  • Add full_feature_name field for databricks.sdk.service.ml.PublishSpec
Read the original announcement →

https://github.com/databricks/databricks-sdk-py/releases/tag/v0.135.0

Related releases