Core Data Custom Store
summary
Explains when to build a custom Core Data store and what actually gets implemented: NSAtomicStore loads and saves an in‑memory cache via load/save/newCacheNode so small datasets (Dashlane’s encrypted passwords and cards) fit and use SQLCipher as the backing DB, while NSIncrementalStore implements execute and newValuesForObject to fetch IDs and fill faults incrementally, trading much higher implementation cost for tunable performance; documentation is flaky but the approaches do work in production.
AI-written summary · about
chapters (14)
- 0:00Core Data custom stores intro
- 1:05Why build a custom store type
- 2:31Atomic vs incremental stores
- 3:59Atomic stores and cache nodes
- 7:35NSIncrementalStore methods & faults
- 11:58Pros and cons of custom stores
- 13:23Custom store learning resources
- 14:10Q&A: Replacing Core Data
- 15:23Q&A: Dashlane’s encrypted store
- 16:06Q&A: REST API custom store
- 17:10Q&A: Propagating object updates
- 18:15Q&A: Core Data threading & container
- 19:45Q&A: Avoid sync data fetching
- 20:30Q&A: Writing data via context
related talks
23mWhat's an App Without Data?
41mUsing Core Data in a Modern SwiftUI Application
06mMeta-Data Caching and Record Store Scalability
46mKEEPING UP IN A DATA-INTENSIVE WORLD WITH SINGLESTORE
47mThe future of Real-time/offline/data
26mEffective Core Data with Swift
59mIndex Internals: RowStore, ColumnStore, and NoStore!
40mBoost App Reliability: Offline Storage Solutions with AWS Expertise