starburst galaxy’s materialized views (using apache iceberg)

join me on a quick test drive of the features of materialized views in starburst galaxy (saas offering powered by trino) which use apache iceberg for persistence and features some pretty cool features around snapshots and awareness of stale data

eliminate rollup’s null confusion (hint: grouping keyword)

rollup functions,such as cube, identify the rolled up totals by using null for the column they are representing a total for – this gets rather confusing when the column itself has null values in the individual rows – this post will show you how to definitively differentiate between these two cases

hive, trino & spark features (their journeys to sql, performance & durability)

different big data sql engines are created to solve a particular lack of focus from existing ones, but sooner or later they all start looking like each other from their list of features and observable behaviors