apollo-caching-strategies

Fetch policies determine Apollo's data retrieval strategy. Set fetchPolicy on queries to cache-first (use cache, fetch if missing), cache-only (cache only), cache-and-network (both simultaneously), network-only (always fetch), no-cache (never cache), or standby (inactive). Each policy balances freshness against network overhead.

Optimistic UI updates render changes instantly before the server responds. Configure the mutation's optimisticResponse option with predicted data, then Apollo merges the mutation result into cache when it arrives. This eliminates perceived latency and improves perceived responsiveness for user actions.

Normalization stores entities by type and ID instead of nested hierarchies. Configure InMemoryCache with typePolicies to define how entities are identified and merged. Use read and write functions to customize field behavior, preventing duplicate data and enabling consistent updates across queries.

Yes. Use the update function in mutations to manually write cache changes, or configure typePolicies with merge functions for automatic normalization. This avoids redundant network requests and keeps UI in sync without triggering fresh queries.

Stale data results from aggressive caching. Use cache-and-network to refresh background data, or cache-first with selective network-only queries for critical updates. Pair with cache expiration logic in typePolicies to evict old data and balance instant responses against data currency.

Yes. Use typePolicies with custom merge functions to concatenate paginated results, or cache-and-network to fetch new pages while displaying cached ones. Implement offset or cursor-based pagination in fetch policies to manage data volume and prevent memory bloat.