Example: 100k-Row Export

Large exports are what the Fast stream path is built for: 100k rows run in a Worker in the browser (the main thread only does one structured clone), with Fast stream finishing in ~0.8s.

Mock data preview

Mock · 3 rows · seed 42

Order ID	Date	Region	Product	Channel	Qty	Unit Price	Amount	Status
ORD-000001	2026-07-21	华东	USB-C 扩展坞	线上	18	789.44	14209.92	pending
ORD-000002	2026-07-28	西南	显示器支架	线上	5	1127.47	5637.35	paid
ORD-000003	2026-07-15	东北	机械键盘	线下	1	898.61	898.61	paid

Implementation

import { exportExcel } from "@marcusok/excel-exporter";

// rows: sales data from your business layer, fields matching the columns
// below (fetching is omitted here; the scenario uses 100k rows — the live
// demo generates mock data for the same sales scenario, with more columns)

const result = await exportExcel({
  filename: "large-export-100k",
  sheets: [
    {
      name: "Sales",
      columns: [
        { prop: "orderId", label: "Order ID", width: 18 },
        { prop: "date", label: "Date", width: 12 },
        { prop: "amount", label: "Amount", width: 14 },
        { prop: "status", label: "Status", width: 10 },
      ],
      data: rows,
    },
  ],
  mode: "auto", // ≥ 50k rows -> worker + stream
  onProgress: (p) => setProgress(p),
});

console.log(result); // engine: "modern-xlsx", mode: "stream", rowCount: 100000

Notes

• At 100k rows auto picks worker + Fast stream: ~0.8s measured (vs 17.5s on the Workbook path);
• The stream path supports multi-row headers and merges, but does not support styles/width/freeze/filter; a console warning is expected;
• onProgress reports every 1000 rows;
• Set decimals explicitly on numeric columns for value consistency with the Workbook path.

Compare auto vs main at 100,000 rows in the play.