What is the difference between Leap partition and Follow partition in data partitioning?

Leap partitioning and Follow partitioning are two strategies for organizing data in a lakehouse or data warehouse. Leap partitioning pre‑creates partitions for future time windows (e.g., the next month or quarter) so that when data arrives, it can be written directly to an existing partition. This reduces the overhead of creating new partitions on the fly and can improve write performance for workloads with predictable time ranges. Follow partitioning, by contrast, creates partitions only when data for a new period arrives; it “follows” the ingestion timeline. This is simpler to set up and works well for unpredictable or streaming workloads where you don’t know the future volume. The choice depends on write patterns, query latency, and maintenance overhead. If you have a slowly changing dimension or a time series with a known horizon, Leap can give you faster writes and fewer partition churn events. If your data arrives irregularly or you need to keep the schema flexible, Follow is usually the safer option. In practice, many teams use a hybrid approach: Leap for the most heavily queried dimensions and Follow for the rest.