Patrick Brosset

Reading large amounts of data from IndexedDB can be slow. The current best approach — batch reading using getAll and getAllKeys with IDBKeyRange — requires two separate requests per batch and doesn't support reverse-order reads. A proposal from the Microsoft Edge team introduces a new getAllRecords() method for IDBObjectStore and IDBIndex that combines keys and values in a single request, supports directional (including reverse) reads, and reduces round-trips to the IDB engine. Benchmark results show roughly 33% faster forward reads and ~68% faster reverse reads compared to existing approaches. A prototype is available in Chrome Canary and Edge Canary behind a feature flag, and the team is seeking developer feedback before pursuing cross-browser standardization.

Even faster IndexedDB reads with getAllRecords