Query decomposition
Query decomposition is the technique of dividing a complex information need into smaller subqueries or subtasks that can be processed separately.
Retrieval Augmented Generation, The Foundational Ideas
Query decomposition is the technique of splitting a complex information need into smaller sub-queries that can be searched or answered separately.
- How it is used
- Splits multi-hop or fragmented information needs into focused retrieval tasks so a RAG system can assemble a more complete answer.
- How deep it goes
- Details atomic queries, per-query summaries, coverage checks, and follow-up retrieval, but does not specify a query-generation algorithm.
- Examples
- Uses UK–EU compliance comparisons, CVaR and IRRBB searches, and a merchant-account procedure spread across documents.
- What you can do after
- Apply it via LevelRAG’s High-Level Searcher or conditional sub-query generation, then check coverage and latency against a simpler retrieval pass.
Read how Retrieval Augmented Generation, The Foundational Ideas explains query decomposition →
Introduction to Generative AI, Second Edition
Decomposition prompting is a technique that breaks a complex problem into smaller subtasks handled in sequence.
- How it is used
- Guides language models through complex documents, planning problems, and educational or tutoring tasks by sequencing subtasks.
- How deep it goes
- Defines sequential decomposition and contrasts it with least-to-most prompting, but stops short of implementation details.
- Examples
- Uses financial-report summarization and stepwise math problem solving as concrete examples.
- What you can do after
- Apply decomposition prompting by turning a complex task into an ordered sequence of smaller subtasks.
Read how Introduction to Generative AI, Second Edition explains query decomposition →
Build an Advanced RAG Application (From Scratch)
Query decomposition transforms a compound information need into multiple atomic sub-queries that can be processed independently.
- How it is used
- Deploys query decomposition as a pre-retrieval transformation in Enterprise RAG, routing and retrieving each distinct information need independently.
- How deep it goes
- Explains embedding dilution, atomic splitting, independent routing and retrieval, synthesis, and simple-query pass-through.
- Examples
- Uses an Uber revenue-and-growth comparison, separate Uber and Lyft filings, and a single-query pass-through case.
- What you can do after
- Build it from scratch with Python and an OpenAI GPT-4o call that returns validated JSON sub-queries.
Read how Build an Advanced RAG Application (From Scratch) explains query decomposition →
What to know first
- retrieval-augmented generationRetrieval-augmented generation is a technique that retrieves relevant information from external sources and supplies it as context to a generative model when producing an output.
- routingRouting is the process of directing an item, request, or task to a destination or processing path according to relevant rules or characteristics.
Related concepts
- decompositionkind of · Decomposition is the process of dividing a complex problem, task, or system into smaller constituent parts that can be handled or analyzed separately.
- atomic queryproduces · An atomic query is an individual information request representing one distinct need extracted from a compound query and answerable independently.
- compound queryaddresses · A compound query combines multiple conditions, subqueries, or information needs to express complex search or retrieval criteria.
- LevelRAGimplemented by · LevelRAG is a retrieval-augmented generation technique that decomposes queries, routes sub-queries to specialized retrievers, and verifies combined evidence before completing generation.
Questions about query decomposition
What is query decomposition?
Query decomposition is the technique of dividing a complex information need into smaller subqueries or subtasks that can be processed separately.
Which Manning books explain query decomposition?
Retrieval Augmented Generation, The Foundational Ideas, Introduction to Generative AI, Second Edition and Build an Advanced RAG Application (From Scratch).
What should I know before learning query decomposition?
Manning books that teach query decomposition usually introduce retrieval-augmented generation and routing first.
What is query decomposition a kind of?
Query decomposition is a kind of decomposition. Decomposition is the process of dividing a complex problem, task, or system into smaller constituent parts that can be handled or analyzed separately.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.


