Document Database
Subject: Computer and Technology noun /ˈdɑːk.jə.mənt ˈdeɪ.təˌbeɪs/ 1 definition
- database-systems
- nosql-databases
- data-modeling
- data-storage
-
Definition (noun)
Entry #235 · UpdatedA Document Database (also called a document-oriented database or document store) is a non-relational database system that stores, indexes, and queries semi-structured information as self-describing documents rather than rows in fixed tables. Each document encapsulates a complete record—using hierarchical formats such as JSON, BSON, or XML—with nested fields, arrays, and embedded subdocuments organized into logical collections. Unlike relational database management systems, a document database does not require a rigid, predefined global schema or cross-table normalization, allowing records in the same collection to vary in structure while keeping related data co-located for fast single-document reads and horizontal scaling.
Common Types
JSON and BSON Document Stores: Purpose-built engines (such as MongoDB and Apache CouchDB) that persist JSON or binary-encoded BSON trees, supporting rich path queries and secondary indexes on nested fields.
XML Document Databases: Specialized repositories (such as BaseX and MarkLogic) built around hierarchical XML documents, offering native XPath and XQuery execution for text-heavy publishing and archival workloads.
Multi-Model Document Databases: Platforms (such as ArangoDB and Couchbase) that integrate document storage with graph traversal, key-value access, or full-text search within a single query engine.
Embedded and Offline-First Document Stores: Lightweight, in-process databases (such as PouchDB and LiteDB) that persist documents directly on mobile or edge devices and synchronize changes with remote clusters.
Example
An e-commerce platform stores each product in a
productscollection as a single BSON document, allowing a laptop entry to holdcpuandram_gbfields while a T-shirt entry in the same collection holds an array ofsizesandcolorvariants. When the storefront renders a product page, the application fetches the entire item—including embedded customer reviews and pricing metadata—in one query without joining multiple normalized tables.