Cube Core v1.6 — Cube Store upgrade, multi-stage pre-aggregations
Blog post from Cube
Cube Core version 1.6 introduces significant enhancements, notably with the upgraded Cube Store that now supports pre-aggregations for multi-stage calculations using the Tesseract data modeling engine. This release includes breaking changes such as strict validation of join paths and access policy matching, emphasizing the need for careful review before deployment. The Cube Store upgrade improves performance by leveraging Apache Parquet, Apache Arrow, and Apache DataFusion, now supporting precise decimals with up to 38 digits of precision and enabling query planning with common-table expressions (CTEs). The release also marks the general availability of query pushdown in the SQL API and introduces a new environment variable for setting default query time zones. Deprecated features include the Elasticsearch driver and support for Vue version 2, while performance optimizations have been made to enhance data model compilation and API instance startup. Users are encouraged to test the changes in a staging environment and report any issues via GitHub and the Slack community.