TabularJS vs SheetJS (xlsx)
SheetJS (the xlsx package) is the de-facto JavaScript Excel library, but it is also large, has a dual-license model, and exposes a low-level workbook/worksheet API that most applications end up wrapping. TabularJS offers the same core use case (reading a spreadsheet file into JSON) behind a single async call, with minimal dependencies and a unified output across 16+ formats.
At a glance
| Feature | TabularJS | SheetJS (xlsx) |
|---|---|---|
| Dependencies | 2 (JSZip, HTML table parser) | Bundled CFB + codepage + others |
| License | MIT | Apache 2.0 (Pro tier commercial) |
| Latest release on npm | Yes (tabularjs) | No: npm stops at 0.18.5; newer releases are on the SheetJS CDN (details) |
| Formats read | 16+ (XLSX, XLS, ODS, CSV, HTML, DBF, SYLK, DIF, Lotus) | Many, depending on build |
| API surface | Single await tabularjs(file) call | workbook → sheet → cell helpers |
| Output shape | Worksheet-first JSON, drop-in for Jspreadsheet | Workbook object, needs utils to convert |
| Write / mutate workbooks | Read-only | Yes |
| Runs in Node & browser | Yes | Yes |
Use SheetJS if you need to write workbooks, author formulas programmatically or do complex workbook manipulation. Use TabularJS if you only need to read files into JSON; that is the overwhelming majority of upload and ETL paths.
SheetJS on npm and npm audit
The xlsx package on the npm registry stops at 0.18.5 (March 2022). SheetJS publishes newer releases only on its own CDN, so npm audit reports two high-severity advisories for it, CVE-2023-30533 (prototype pollution, fixed in 0.19.3) and CVE-2024-22363 (ReDoS, fixed in 0.20.2), with “No fix available”. Both concern reading crafted files. If you keep SheetJS, installing 0.20.3 fromcdn.sheetjs.com clears both.
What the advisories mean and how to fix npm audit
Migration: Node.js
A typical SheetJS read loop looks like this:
import * as XLSX from 'xlsx';
import fs from 'fs';
const buffer = fs.readFileSync('./sales.xlsx');
const workbook = XLSX.read(buffer, { type: 'buffer' });
const sheetName = workbook.SheetNames[0];
const worksheet = workbook.Sheets[sheetName];
const rows = XLSX.utils.sheet_to_json(worksheet, { header: 1 });
console.log(rows);The same thing with TabularJS is a single call:
import tabularjs from 'tabularjs';
const result = await tabularjs('./sales.xlsx');
console.log(result.worksheets[0].data);TabularJS detects the format from the path or file contents, so the same call works for .xlsx, .xls, .ods, .csv,.html and the rest.
Migration: browser upload
// With SheetJS
input.addEventListener('change', async (e) => {
const data = await e.target.files[0].arrayBuffer();
const workbook = XLSX.read(data);
const rows = XLSX.utils.sheet_to_json(
workbook.Sheets[workbook.SheetNames[0]],
{ header: 1 }
);
render(rows);
});// With TabularJS
input.addEventListener('change', async (e) => {
const result = await tabularjs(e.target.files[0]);
render(result.worksheets[0].data);
});No ArrayBuffer conversion, no sheet-name lookup, no utils.sheet_to_json.
Migration: rendering in Jspreadsheet
With SheetJS you typically convert cells, then re-map them into Jspreadsheet'sdata/columns shape. TabularJS returns that shape directly:
import jspreadsheet from 'jspreadsheet-ce';
import tabularjs from 'tabularjs';
input.addEventListener('change', async (e) => {
const result = await tabularjs(e.target.files[0]);
// The shape matches jspreadsheet() 1:1, no mapping required
jspreadsheet(target, result);
});What you keep
- Formulas: preserved for XLSX, XLS, ODS, XML and SYLK.
- Merged cells and basic styles where the source supports them.
- A single JSON shape, identical across every format.
Ready to switch?
Install TabularJS and replace your SheetJS read calls in a single commit.