xlsx-fast-node

High-performance Excel (.xlsx) parser with a native Rust core.

Features

• Fast: SIMD-optimized cell reference parsing (NEON on ARM, AVX2/SSE2 on x86)
• Memory Efficient: Memory-mapped files and streaming decompression
• Complete: Handles formulas, dates (1900/1904 systems), shared strings
• Safe: Rust's memory safety guarantees
• NDJSON Streaming: For processing huge files without loading into memory

Installation

npm install xlsx-fast-node

Quick Start

const { parseXlsx, parseXlsxNdjson } = require('xlsx-fast-node');

// Parse entire workbook
const workbook = parseXlsx('/path/to/file.xlsx');
console.log(`Sheets: ${workbook.sheets.length}`);

for (const sheet of workbook.sheets) {
  console.log(`Sheet: ${sheet.name} (${sheet.cells.length} cells)`);
  for (const cell of sheet.cells) {
    console.log(`  [${cell.row},${cell.col}] = ${JSON.stringify(cell.value)}`);
  }
}

Options

const workbook = parseXlsx('/path/to/file.xlsx', {
  decodeDates: true,       // Convert Excel dates to ISO8601 (default: true)
  includeFormulas: false,  // Include formula text (default: false)
  rawValues: false,        // Skip numeric/date parsing (default: false)
  maxInflate: 128 * 1024 * 1024,  // Max decompressed size in bytes (default: 128 MiB)
});

NDJSON Streaming

For large files, stream to NDJSON without loading everything in memory:

parseXlsxNdjson('/path/to/large.xlsx', '/tmp/out.ndjson');

The NDJSON output format (one JSON object per line):

{"kind":"sheet","id":1,"name":"Sheet1"}
{"kind":"cell","sheet_id":1,"row":1,"col":1,"type":"string","v":"Hello"}
{"kind":"cell","sheet_id":1,"row":1,"col":2,"type":"number","v":42}
{"kind":"cell","sheet_id":1,"row":1,"col":3,"type":"date","iso":"2024-01-15","serial":45276}

Data Types

CellValue

Cells can have one of these value types:

• Empty - Empty cell
• Bool { v: boolean } - Boolean value
• Number { v: number } - Numeric value
• String { v: string } - String value (from shared strings)
• Raw { v: string } - Raw string value (inline)
• Error { v: string } - Error value
• Date { iso: string, serial: number } - Date value (ISO8601 string + Excel serial)

TypeScript Support

TypeScript definitions are included:

import { parseXlsx, ParseOptions, Workbook, Sheet, Cell } from 'xlsx-fast-node';

const workbook: Workbook = parseXlsx('data.xlsx');

Performance

Benchmarks show significant performance improvements over pure JavaScript/Node.js alternatives:

• 10x faster than openpyxl (Python)
• 5x faster than SheetJS in streaming mode
• Uses 50% less memory for large files

Platform Support

Prebuilt binaries available for:

• macOS (Intel & Apple Silicon)
• Linux (x64, ARM64)
• Windows (x64)

License

MIT

Related Packages

• docx-fast-node - Word (.docx) parser
• pptx-fast-node - PowerPoint (.pptx) parser