Why Streaming?

When JSON files exceed memory limits (e.g., 1GB+), JSON.parse() causes OOM (Out of Memory). Streaming allows processing nodes one by one with low memory footprint.

Common Solutions

1. NDJSON (Newline Delimited JSON)

One JSON object per line, naturally suited for streaming.

{"id": 1, "data": "..."}
{"id": 2, "data": "..."}

2. Streaming Libraries (e.g., stream-json)

Event-driven parsing for single large arrays or objects.

const { parser } = require('stream-json');
const { streamArray } = require('stream-json/streamers/StreamArray');

fs.createReadStream('large.json')
  .pipe(parser())
  .pipe(streamArray())
  .on('data', ({ value }) => process(value));

Browser-side Streaming

Leverage ReadableStream and TextDecoder, combined with WebAssembly or custom parsers, to handle large file imports on the client.

Summary

Streaming is essential for heavy JSON applications. Choosing the right format (like NDJSON) is often more effective than optimizing parsing algorithms.