strawberry.dataloader.DataLoader automatically deduplicates keys: duplicate load() calls don't cause extra batch_load_fn invocations
Strawberry GraphQL's strawberry.dataloader.DataLoader handles duplicate keys automatically without custom logic. When you call loader.load('a'), loader.load('b'), loader.load('a'), loader.load('c'), loader.load('b'), the DataLoader automatically deduplicates the keys before calling batch_load_fn, so batch_load_fn receives ['a', 'b', 'c'] (3 unique keys), not the 5 original calls. The loader then maps the batch results back to all 5 callers correctly. This means you don't need custom deduplication logic in your batch function or a wrapper that tracks seen keys. Just implement batch_load_fn to return results in the same order as the input keys, and DataLoader handles the rest. This is the standard DataLoader pattern from Facebook's dataloader spec that Strawberry implements. Verification: create a DataLoader with a batch function that prints its input, call load() with duplicates, and observe that batch_load_fn only sees unique keys once.