hologres-uv-compute

To compute real-time UV and PV with deduplication, use Dynamic Tables with incremental refresh and the RoaringBitmap extension. This approach supports incremental pipelines on billions of events for scalable analytics.

RoaringBitmap handles cross-day UV aggregation using the RB_OR_AGG function. This enables exact deduplicated user counts and flexible time-range analysis across multiple days with low latency.

Yes, real-time UV computation requires a robust UID-mapping strategy when using text user IDs. You can encode text UIDs into integers using the hg_id_encoding_int4 function for RoaringBitmap processing.

The best way to set up incremental deduplication for event-level analytics is creating source and dynamic tables, enabling the RoaringBitmap extension, and running incremental refreshes. This supports retention studies and cross-region campaigns.

Hologres supports exact UV counts without repeated raw log scanning by leveraging Dynamic Tables and RoaringBitmap. This architecture maintains deduplicated user IDs incrementally for scalable, low-latency queries.