Lazy vs eager loading

Lazy loading fetches related data when first accessed. Eager loading (join, include, preload) fetches relationships upfront. Choice depends on request shape, cardinality, serialization, and memory.

| Strategy | Pays off when | Risk | |----------|---------------|------| | Lazy | Relationship rarely used | N+1, queries after transaction closed | | Eager | Relationship always needed in response | Over-fetch, wide joins, memory |

On interviews: connect loading strategy to N+1 prevention, API DTO shape, and boundaries between persistence models and response objects.

Common pitfalls: lazy loading during JSON serialization; eager loading broad graphs nobody uses; loading inside loops without batching.

The trade-off is query count and latency versus memory and response size.

Checklist:

• Know relationship cardinality.
• Measure query count per endpoint.
• Map loaded data to response needs.
• Prefer explicit includes over implicit lazy access.