ndjson

NDJSON, short for newline-delimited JSON, is a data interchange format in which each line of a text file is a complete JSON value. Typically each line contains a JSON object, but any valid JSON value is allowed. Lines are delimited by standard newline characters, and there is no outer JSON array or separator.

Unlike a single JSON document, NDJSON represents a stream of independent records that can be produced and consumed incrementally. This makes it suitable for log files and data pipelines, where records are appended over time or processed in parallel. Each line must be a standalone, valid JSON value.

Common use cases include streaming ingestion into data stores, log aggregation, and data interchange between services. NDJSON is popular in environments that require line-oriented processing, such as search engines and big data systems, where parsers can deserialize records one at a time without loading the entire dataset into memory.

Advantages include low memory footprint, easy incremental parsing, and compatibility with standard JSON libraries. Limitations include the absence of a single, verifiable JSON document for the whole dataset, lack of formal standardization, and the potential need for separate handling of line endings and schema evolution.

NDJSON is widely supported by many tools and languages, and is often used alongside JSON Lines. It is compatible with systems that process text streams and can be ingested by bulk imports, log shippers, and ETL pipelines.