A blockchain indexer can return data that later disappears from the canonical chain. The original article, published on November 13, 2026, explains why a chain reorganization is a data-consistency problem for subgraphs, indexers and the APIs built on them—not just a reason to delete an invalid transaction.
Consider an indexed transfer from Alice to Bob that is later removed by a reorganization. Removing its transaction record alone would leave other data derived from it out of sync. The indexer must also roll back changes to related entities, balances, total value locked (TVL) and aggregate statistics, then process the canonical chain again.
The article distinguishes freshness from finality. Data served at latest may be more current but more exposed to reorganization; safe and finalized represent different trade-offs between currency and confidence. Consistency also extends beyond the indexer: API caches, downstream business state and notifications that retract earlier events need to reflect a rollback.
When assessing an indexing service, the article recommends checking whether it can detect the canonical chain, roll back state, reprocess data, provide consistent snapshots, recover from deep reorganizations and make those processes observable.
Original article · Published November 13, 2026, at 02:26 (UTC+08:00).