JSON to NDJSON - Generate Streaming JSON

Convert JSON arrays to NDJSON format for streaming, BigQuery imports, and Elasticsearch. Browser-based, secure, supports LF/CRLF line endings.

JSON Input

Use CRLF (Windows)

NDJSON Output

0 objects

Tip: NDJSON is perfect for streaming data to BigQuery, Elasticsearch bulk API, log aggregation, and parallel processing. Each line is a complete JSON object.

YAML to CSV - Export Config Data to Spre…

Properties to JSON - Java Config Migrati…

JSON to Plist Converter - Generate Apple…

CSV to XML Generator for Legacy Systems

XML to YAML - Migrate Configs to Modern …

View all Data serializers →

Frequently Asked Questions

NDJSON (Newline Delimited JSON) is a format where each line contains a valid JSON object, separated by newline characters. Unlike standard JSON arrays, NDJSON allows streaming processing without loading the entire dataset into memory. Each line is independent and can be parsed separately, making it ideal for large datasets and real-time data processing.

NDJSON offers several advantages: streaming capabilities for processing large datasets without loading everything into memory, easier log aggregation since each entry is self-contained, better performance for append operations (just add a new line), simpler parallel processing since lines are independent, and fault tolerance where corrupt lines don't invalidate the entire file.

BigQuery and Elasticsearch natively support NDJSON for bulk imports. In BigQuery, you can use bq load with --source_format=NEWLINE_DELIMITED_JSON to import data efficiently. Elasticsearch's bulk API expects NDJSON for indexing multiple documents at once. This format allows both services to stream-process imports without memory constraints, making it ideal for importing millions of records.

For converting multiple objects to NDJSON, yes, your input should be a JSON array. The converter will extract each element and output it as a separate line. However, if you input a single JSON object (not in an array), the tool will output it as a single NDJSON line. You can also paste multiple JSON objects on separate lines, and they'll be preserved as valid NDJSON.

Nested objects and arrays within each top-level element are preserved exactly as-is on each NDJSON line. For example, if your JSON array contains objects with nested arrays or complex structures, each object will be serialized as a complete, valid JSON document on its own line. Only the outer array structure is removed.

Use LF (Line Feed, \n) for Unix/Linux systems, cloud services like BigQuery and Elasticsearch, and modern data processing tools. Use CRLF (Carriage Return + Line Feed, \r\n) only if you're working on Windows and need compatibility with legacy Windows applications. Most streaming JSON processors expect LF, which is the default and recommended choice.

NDJSON typically compresses very well with gzip or similar algorithms, often better than equivalent JSON arrays. Because each line has similar structure (for homogeneous data), compression algorithms can efficiently identify patterns. Additionally, NDJSON files can be compressed and decompressed in streams without loading the entire file, making it more memory-efficient for large datasets.

This browser-based tool can handle JSON files limited by your browser's available memory (typically up to several hundred megabytes). For very large files (gigabytes), consider using command-line tools like jq (jq -c '.[]') or streaming parsers in Python/Node.js. However, for most web-based workflows and datasets under 100MB, this tool provides a convenient, secure solution without server uploads.

Each line in NDJSON must be valid JSON. You can validate by: parsing each line individually with JSON.parse() in JavaScript, using command-line tools like 'jq -e' which will error on invalid JSON, or testing imports in your target system (BigQuery, Elasticsearch). The converter ensures each output line is properly formatted and doesn't add trailing commas or array brackets.

NDJSON is widely supported: BigQuery (Google Cloud), Elasticsearch and OpenSearch, MongoDB (mongoimport), Apache Kafka, Logstash and Fluentd, jq command-line tool, pandas (read_json with lines=True), Apache Spark, Snowflake, and most modern ETL and streaming platforms. It's become a de facto standard for streaming and bulk data operations.

JSON Input

NDJSON Output

Related Tools