Cache BigQuery table definitions in Python SDK#39028
Conversation
Summary of ChangesHello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request optimizes BigQuery interactions in the Python SDK by implementing a class-level cache for table definitions. By caching these definitions, multiple wrapper instances within the same worker process can reuse existing metadata, significantly reducing the number of redundant 'tables.get' API requests. The implementation includes robust cache management with size limits, expiration policies, and lifecycle-based invalidation to maintain accuracy. Highlights
New Features🧠 You can now enable Memory (public preview) to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize the Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counterproductive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces an in-memory cache for BigQuery table definitions within BigQueryWrapper to reduce redundant API calls, complete with TTL expiration, eviction, and thread safety. The review feedback suggests optimizing the cache-clearing logic by using clear() when emptying the entire cache, and reducing the default TTL from 1 hour to 5 minutes to prevent workers from using stale metadata in streaming or long-running pipelines.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| if (project_id is not None and dataset_id is not None and | ||
| table_id is not None): | ||
| cache_key = cls._table_definition_cache_key( | ||
| project_id, dataset_id, table_id) | ||
| cls._table_definition_cache.pop(cache_key, None) | ||
| return | ||
|
|
There was a problem hiding this comment.
When clearing the entire cache (i.e., when project_id, dataset_id, and table_id are all None), we can use cls._table_definition_cache.clear() instead of building a list of all keys and popping them one by one. This is much more efficient and simpler.
with cls._table_definition_cache_lock:
if project_id is None and dataset_id is None and table_id is None:
cls._table_definition_cache.clear()
return
if (project_id is not None and dataset_id is not None and
table_id is not None):
cache_key = cls._table_definition_cache_key(
project_id, dataset_id, table_id)
cls._table_definition_cache.pop(cache_key, None)
return|
|
||
| # Shared by wrapper instances within one Python SDK worker process. | ||
| _TABLE_DEFINITION_CACHE_MAX_ENTRIES = 256 | ||
| _TABLE_DEFINITION_CACHE_TTL_SECS = 60 * 60 |
There was a problem hiding this comment.
A TTL of 1 hour (60 * 60 seconds) is quite long for caching table definitions. In streaming pipelines or long-running jobs where table schemas might be updated dynamically (e.g., adding a nullable column), workers could use stale metadata for up to an hour, leading to schema mismatch errors. Consider reducing the default TTL to a more conservative value (e.g., 5 or 10 minutes) or making it configurable via pipeline options.
| _TABLE_DEFINITION_CACHE_TTL_SECS = 60 * 60 | |
| _TABLE_DEFINITION_CACHE_TTL_SECS = 5 * 60 |
|
Assigning reviewers: R: @shunping for label python. Note: If you would like to opt out of this review, comment Available commands:
The PR bot will only process comments in the main thread (not review comments). |
This adds a small class level cache around BigQueryWrapper.get_table() so multiple wrapper instances in the same worker can reuse table definitions instead of repeatedly calling BigQuery tables.get.
The cache is bounded, expires entries after a conservative TTL, and is invalidated when Beam creates or deletes a table.
addresses #34076