Super Table (ST) is an initiative at our company aimed at defining, building, and sharing high-quality data products. These products come with formal ownership and commitments.
Super Tables are designed as pre-computed, denormalized, and consistently consolidated attributes and insights of entities or events. They are optimized for common and efficient analytic use cases.
Super Tables feature well-defined service level agreements (SLAs) and are instrumental in simplifying data discovery and downstream data processing.
Additionally, Super Tables carry enterprise-grade commitments crucial for critical use cases. These include high data quality, availability (ensuring users can rely on STs), disaster recovery, proper documentation, maintenance, and governance.
Interestingly, Super Tables align seamlessly with Data Mesh, primarily because Data-as-a-product (or Data Product) is a core principle. The Super Tables are geared towards developing top-tier data products with clearly defined ownership and responsibilities.
A noteworthy comment from Jon Cook on Cliff Leung on LinkedIn: Super Tables suggests that this concept is akin to older Materialized Views. The idea is to create “specific data structures per product but allow these to be curated, annotated, and published for reuse if desired”. I generally concur, even though I haven’t read the entire article yet.
Similar to One Big Table according to This Tweet. Personally, I align with the OBT approach, which is akin to the super tables concept. I’m not a big fan of Kimball/dimensional modeling, but this perspective is worth considering: Atheer Alabdullatif on Twitter: “https://t.co/CNAv1VoPUu” / Twitter