pagination-with-preserved-aggregates

Summary statistics break because the same paginated query is incorrectly used to compute totals that should cover the entire filtered dataset. Separating paged row fetching from full-dataset aggregate queries keeps stats accurate across all pagination states.

Implement a two-query backend pattern that separates paged row fetching from full-dataset aggregate queries. This ensures your Excel and CSV exports always reflect the complete filtered dataset rather than just the current page.

Yes, this server-side pagination approach works with Prisma, Drizzle, and Sequelize, as well as backend frameworks like Elysia, Express, and Fastify. It applies to any data-table interface with summary cards and export functionality.

Prevent double-fetch bugs by implementing frontend state management that ensures stats cards read totals directly from the server, not the current page's row count. This avoids redundant API calls and keeps summary data consistent.

Use the included reusable smoke test template to verify the core invariant that totals remain consistent across page 1, page 2, full exports, and group-by modes. This enforces that your summary statistics never drift during pagination.