sheetstream

sheetstream reads and writes XLSX and CSV from Node.js without holding the rows in JavaScript. A Rust core does the parsing and zipping and rows cross in batches, so a million-row export fits in a small serverless function. It takes any iterable or async iterable as a source, so a database cursor or a generator works as is.

How it works

1. 01

   Native Rust code (through napi-rs) does the XML, zip and CSV work, so the rows never become JavaScript objects in bulk.

2. 02

   Rows go in and out in batches of 1000 by default, and a parser thread runs a few batches ahead of your loop.

3. 03

   Writing in constant mode streams the sheet XML to a temp file and compresses it into the output, so the cost is disk, not RAM.

What it does

• writeXlsx takes any Iterable or AsyncIterable of arrays or objects, and readXlsx is an async iterator of batches
• Streaming CSV read and write, plus csvStream, a Node Readable with real backpressure for HTTP responses
• XLSX formatting: header style, column widths, per-column number formats, frozen header row and header filter, at no extra memory
• Multiple sheets, choose by name or index, with xlsxStream and toBuffer for HTTP and small outputs
• Prebuilt binaries for macOS (arm64, x64), Linux (x64 glibc, x64 musl, arm64 glibc) and Windows (x64), nothing to compile on install