v1.7.15
This release includes bug fixes for aggregating index type casting, rolling window schema preservation, and grain projection, along with features for query row limits and record batch splitting to improve query handling.
1.7.15 (2026-07-30)
Bug Fixes
- cubestore: cast aggregating index aggregation results back to the declared column types (#11413) (26c4801)
- cubestore: keep the projection schema when rewriting a rolling window (#11410) (1bf2dd8)
- tesseract: project the requested grain into rolling-window CTEs (#11400) (80d1159), closes #11410 #11413
Features
Source: original entry ↗
More from Cube
Follow Cube to get its new changes in your feed and email digest.
Load partitions for multi-stage rolling window in tesseract
Fixed an issue where the tesseract engine was not properly loading partitions for multi-stage rolling window operations, which could cause incorrect behavior or failures.
v1.8.0
Version 1.8.0 includes bug fixes for database drivers and schema compilation, new stricter validation for pre-aggregation queries, primary key detection in data model generation, and performance optimizations to reduce memory usage and improve result parsing efficiency.
Bound multi-stage stages and stop pre-aggregation walk doubling per level
Fixed an issue in tesseract where multi-stage stages were not properly bounded and the pre-aggregation walk was doubling per level, affecting query performance and correctness.