概念概述
文档分块策略（Document Chunking Strategies）指的是：分块决定文档在嵌入前如何切分，边界选择决定检索能否返回连贯且自包含的片段，直接影响答案质量。
从更严格的定义看，Chunking is the process of dividing source documents into passages small enough to embed and retrieve, with strategies that balance semantic coherence against index size and precision. 该概念在工程实践中的价值，在于它能帮助团队用统一的语言描述一个具体的质量属性或设计约束，从而在评审与协作中减少歧义。
为什么需要它
It is a design decision in every retrieval pipeline that ingests long or structured documents. 它与 AI Guardrails and Output Validation、Retrieval-Augmented Generation、Vector Databases 等概念经常一起出现，用于解决相近的工程问题。理解它的适用场景，有助于在多种备选方案之间做出有依据的取舍，而不是凭直觉选择技术方案。
常见误解
- a fixed character count is always the best boundary
- larger chunks give strictly better answers
需要避免的反模式
- splitting on arbitrary lengths and severing sentences or tables
- ignoring document structure such as headings and code blocks
实践要点
- Chunks too large dilute the embedding and waste context, while chunks too small lose the surrounding meaning needed to answer, so overlap and structure-aware splits help.
- Respecting document hierarchy, such as headings and sections, produces passages that stand alone and match retrieval queries more reliably.
- Metadata attached to chunks, including source, section, and position, enables filtering and lets the generator cite where an answer came from.