Parent Document Retrieval
A RAG retrieval strategy that first retrieves fine-grained child chunks, then returns the parent document or larger context block to which they belong, in order to provide more complete context.
Timeline (last 90 days)
Parent-document retrieval is a retrieval pattern targeting the chunk-boundary problem: at indexing time small chunks are split for embedding matching, and upon a hit the larger unit to which they belong (the parent document or parent paragraph) is returned.
Parent-document retrieval indexes small child chunks to ensure retrieval precision, but returns to the model the larger parent chunk to which the child chunk belongs, requiring a mapping from child chunks to parent chunks to be maintained.
RAG tutorials fail to make clear that chunk boundaries are the core problem; better chunking strategies include overlapping windows, semantic chunking, and parent document retrieval; the real fix lies in rethinking the content that is stored itself.
Directions for improving the RAG chunking problem include better chunking strategies such as overlapping windows, semantic chunking, and parent document retrieval; the fundamental solution lies in rethinking what is stored, and sometimes a structured representation of the information should be stored rather than the raw text.
All Facts (4)
Parent-document retrieval is a retrieval pattern targeting the chunk-boundary problem: at indexing time small chunks are split for embedding matching, and upon a hit the larger unit to which they belong (the parent document or parent paragraph) is returned.
50%UnverifiedParent-document retrieval indexes small child chunks to ensure retrieval precision, but returns to the model the larger parent chunk to which the child chunk belongs, requiring a mapping from child chunks to parent chunks to be maintained.
50%UnverifiedRAG tutorials fail to make clear that chunk boundaries are the core problem; better chunking strategies include overlapping windows, semantic chunking, and parent document retrieval; the real fix lies in rethinking the content that is stored itself.
50%UnverifiedDirections for improving the RAG chunking problem include better chunking strategies such as overlapping windows, semantic chunking, and parent document retrieval; the fundamental solution lies in rethinking what is stored, and sometimes a structured representation of the information should be stored rather than the raw text.
50%