BigQueryOfflineStore facing problems with complex feature views
- Dominant language
- Python
- Stars
- 7.3k
- Forks
- 1.4k
- Avg merge
- 3d 16h
- Merged PRs (30d)
- 15
Description
## Expected Behavior
I expect BigQueryOffline store would be able to have freedom in joining multiple feature views at large scale, including table-based feature views and SQL query-based feature views
## Current Behavior
When I conducted point-in-time joins across like 30 feature views, BQ client thrown this error
``` bash
400 Resources exceeded during query execution: Not enough resources for query planning - too many subqueries or query is too complex.
```
## Steps to reproduce
1. Prepare 2 table-based feature views
2. Prepare 2 query-based feature views (with pivot statement on 10 values)
3. Conduct point-in-time joining on all feature views
### Specifications
- Version: 0.27.1
- Platform: Linux
## Possible Solution
Convert BQ template from using `WITH` to using `TEMP TABLE`
Contributor guide
Assessment
This issue has not been assessed yet.