How to Parallel Index Scan in YSQL For Temporal Joins

In a recent blog, Mark Peacock discusses the advancements in executing temporal joins in YSQL, particularly emphasizing the benefits of using Parallel Index Scan with YugabyteDB, which now fully supports PostgreSQL's Parallel Query (PQ). Previously, implementing temporal joins required complex workarounds involving bucketized indexes and query rewrites to enable Parallel Append, adding significant complexity. However, with the enhancements in the YSQL optimizer as of version 2025.2.2, these joins can now be executed using a Parallel Index Scan, leveraging PostgreSQL indexes and native SQL, resulting in significant performance gains—up to 50–55% faster than non-PQ plans. This method allows for double parallelism by combining YugabyteDB's tablet-level parallelism with PostgreSQL's PQ execution, optimizing large time-window analytics and HTAP-style reads over range-indexed columns. Configurations to enable these features are done through Global User Configuration (GUC) settings, and future updates aim to make these settings enabled by default, simplifying the process further.