BigQuery Integer Partition Support
BigQuery tables partitioned by integer columns can be written to and merged into, with the integer column selected as the merge predicate.
Some BigQuery workloads organize data using integer partitions instead of dates - think customer segments, numeric ranges, or custom “day” encodings. Until now, Artie’s merge logic focused on time-based partitions, which covered most schemas but didn’t fit teams using integers to keep massive tables fast and organized. This update expands that flexibility without changing how existing BigQuery pipelines work.
Artie can now write to and merge into tables partitioned by integer columns. Whether you’re partitioning by something like customer_id or a numeric transaction_day, you can select that column as your merge predicate when configuring your table. Artie will handle partition-aware merging behind the scenes, keeping performance high even as tables scale.
Why this matters:
- Supports a broader range of BigQuery schema patterns
- Improves merge efficiency for large, integer-partitioned tables
- Enables cleaner modeling for teams that don’t use time-based partitions
- Reduces the need to redesign tables just to fit replication workflows
- Aligns with existing BigQuery best practices for performance tuning