Web Workers

Parsing and writing large spreadsheets takes time. During the process, if the SheetJS library is running in the web browser, the website may freeze.

Workers provide a way to off-load the hard work so that the website does not freeze during processing. The work is still performed locally. No data is sent to a remote server.

The following diagrams show the normal and Web Worker flows when exporting a dataset. The regions with a red background mark when the browser is frozen.

Normal Export	Web Worker Export

Browser Compatibility

IE10+ and modern browsers support basic Web Workers. Some APIs like fetch were added later. Feature testing is highly recommended.

Inline Workers

Due to limitations of the live code blocks, all of the workers in this section are in-line. The code is embedded in template literals. For production sites, typically workers are written in separate JS files.

Example (click to show)

For example, an in-line worker like

  const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', (e) => {
  /* Pass the version string back  */
  postMessage({ version: XLSX.version });
}, false);
  `])));

would typically be stored in a separate JS file like "worker.js":

worker.js
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', (e) => {
  /* Pass the version string back  */
  postMessage({ version: XLSX.version });
}, false);

and the main script would pass a URL:

  const worker = new Worker("./worker.js");

Installation

In all cases, importScripts in a Worker can load the Standalone scripts

importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

For production use, it is highly encouraged to download and host the script.

ECMAScript Module Support (click to hide)

Browser Compatibility

ESM is supported in Web Workers in the Chromium family of browsers (including Chrome and Edge) as well as in browsers powered by WebKit (including Safari).

For legacy browsers like Firefox and IE, importScripts should be used.

Browser ESM imports require a complete URL including the .mjs extension:

import * as XLSX from "https://cdn.sheetjs.com/xlsx-latest/package/xlsx.mjs";

When using Worker ESM, the Worker constructor must set the type option:

const worker = new Worker(
  url_to_worker_script,
  { type: "module" } // second argument to Worker constructor
);

Inline workers additionally require the Blob MIME type text/javascript:

const worker_code = `\
/* load standalone script from CDN */
import * as XLSX from "https://cdn.sheetjs.com/xlsx-latest/package/xlsx.mjs";
// ... do something with XLSX here ...
`;
const worker = new Worker(
  URL.createObjectURL(
    new Blob(
      [ worker_code ],
      { type: "text/javascript" } // second argument to the Blob constructor
    )
  ),
  { type: "module" } // second argument to Worker constructor
);

Live Demos

Downloading a Remote File

fetch in Web Workers

fetch was enabled in Web Workers in Chrome 42 and Safari 10.3

Typically the Web Worker performs the fetch operation, processes the workbook, and sends a final result (HTML table or raw data) to the main browser context:

Live Demo (click to show)

In the following example, the script:

downloads https://sheetjs.com/pres.numbers in a Web Worker
loads the SheetJS library and parses the file in the Worker
generates an HTML string of the first table in the Worker
sends the string to the main browser context
adds the HTML to the page in the main browser context

Result

Live Editor

function SheetJSFetchDLWorker() {
  const [__html, setHTML] = React.useState("");

  return ( <>
    <button onClick={() => {
      /* this mantra embeds the worker source in the function */
      const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', async(e) => {
  try {
    /* Fetch file */
    const res = await fetch("https://sheetjs.com/pres.numbers");
    const ab = await res.arrayBuffer();

    /* Parse file */
    const wb = XLSX.read(ab, {dense: true});
    const ws = wb.Sheets[wb.SheetNames[0]];

    /* Generate HTML */
    const html = XLSX.utils.sheet_to_html(ws);

    /* Reply with result */
    postMessage({ html });
  } catch(e) {
    /* Pass the error message back */
    postMessage({html: String(e.message || e).bold() });
  }
}, false);
      `])));
      /* when the worker sends back the HTML, add it to the DOM */
      worker.onmessage = function(e) { setHTML(e.data.html); };
      /* post a message to the worker */
      worker.postMessage({});
    }}><b>Click to Start</b></button>
    <div dangerouslySetInnerHTML={{ __html }}/>
  </> );
}

function SheetJSFetchDLWorker() {
  const [__html, setHTML] = React.useState("");

  return ( <>
    <button onClick={() => {
      /* this mantra embeds the worker source in the function */
      const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', async(e) => {
  try {
    /* Fetch file */
    const res = await fetch("https://sheetjs.com/pres.numbers");
    const ab = await res.arrayBuffer();

    /* Parse file */
    const wb = XLSX.read(ab, {dense: true});
    const ws = wb.Sheets[wb.SheetNames[0]];

    /* Generate HTML */
    const html = XLSX.utils.sheet_to_html(ws);

    /* Reply with result */
    postMessage({ html });
  } catch(e) {
    /* Pass the error message back */
    postMessage({html: String(e.message || e).bold() });
  }
}, false);
      `])));
      /* when the worker sends back the HTML, add it to the DOM */
      worker.onmessage = function(e) { setHTML(e.data.html); };
      /* post a message to the worker */
      worker.postMessage({});
    }}><b>Click to Start</b></button>
    <div dangerouslySetInnerHTML={{ __html }}/>
  </> );
}

Creating a Local File

Writing files from Web Workers

XLSX.writeFile will not work in Web Workers! Raw file data can be passed from the Web Worker to the main browser context for downloading.

Typically the Web Worker receives an array of JS objects, generates a workbook, and sends a URL to the main browser context for download:

Live Demo (click to show)

In the following example, the script:

sends a dataset (array of JS objects) to the Web Worker
generates a workbook object in the Web Worker
generates a XLSB file using XLSX.write in the Web Worker
generates an object URL in the Web Worker
sends the object URL to the main browser context
performs a download action in the main browser context

Result

Live Editor

function SheetJSWriteFileWorker() {
  const [__html, setHTML] = React.useState("");

  const data = [
    { "SheetJS": "வணக்கம்", "in": "สวัสดี", "Web": "你好", "Workers": "가지마" },
    { "SheetJS": 1, "in": 2, "Web": 3, "Workers": 4 },
  ];

  return ( <>
    <button onClick={() => { setHTML("");
      /* this mantra embeds the worker source in the function */
      const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', async(e) => {
  try {
    /* Create a new workbook from the data */
    const ws = XLSX.utils.json_to_sheet(e.data.data);
    const wb = XLSX.utils.book_new();
    XLSX.utils.book_append_sheet(wb, ws, "Data");

    /* Write XLSB data (Uint8Array) */
    const u8 = XLSX.write(wb, { bookType: "xlsb", type: "buffer" });

    /* Generate URL */
    const url = URL.createObjectURL(new Blob([u8]));

    /* Reply with result */
    postMessage({ url });
  } catch(e) {
    /* Pass the error message back */
    postMessage({error: String(e.message || e).bold() });
  }
}, false);
      `])));
      /* when the worker sends back the data, create a download */
      worker.onmessage = function(e) {
        if(e.data.error) return setHTML(e.data.error);

        /* this mantra is the standard HTML5 download attribute technique */
        const a = document.createElement("a");
        a.download = "SheetJSWriteFileWorker.xlsb";
        a.href = e.data.url;
        document.body.appendChild(a);
        a.click();
        document.body.removeChild(a);
      };
      /* post a message to the worker */
      worker.postMessage({ data });
    }}><b>Click to Start</b></button>
    <div dangerouslySetInnerHTML={{ __html }}/>
  </> );
}

function SheetJSWriteFileWorker() {
  const [__html, setHTML] = React.useState("");

  const data = [
    { "SheetJS": "வணக்கம்", "in": "สวัสดี", "Web": "你好", "Workers": "가지마" },
    { "SheetJS": 1, "in": 2, "Web": 3, "Workers": 4 },
  ];

  return ( <>
    <button onClick={() => { setHTML("");
      /* this mantra embeds the worker source in the function */
      const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', async(e) => {
  try {
    /* Create a new workbook from the data */
    const ws = XLSX.utils.json_to_sheet(e.data.data);
    const wb = XLSX.utils.book_new();
    XLSX.utils.book_append_sheet(wb, ws, "Data");

    /* Write XLSB data (Uint8Array) */
    const u8 = XLSX.write(wb, { bookType: "xlsb", type: "buffer" });

    /* Generate URL */
    const url = URL.createObjectURL(new Blob([u8]));

    /* Reply with result */
    postMessage({ url });
  } catch(e) {
    /* Pass the error message back */
    postMessage({error: String(e.message || e).bold() });
  }
}, false);
      `])));
      /* when the worker sends back the data, create a download */
      worker.onmessage = function(e) {
        if(e.data.error) return setHTML(e.data.error);

        /* this mantra is the standard HTML5 download attribute technique */
        const a = document.createElement("a");
        a.download = "SheetJSWriteFileWorker.xlsb";
        a.href = e.data.url;
        document.body.appendChild(a);
        a.click();
        document.body.removeChild(a);
      };
      /* post a message to the worker */
      worker.postMessage({ data });
    }}><b>Click to Start</b></button>
    <div dangerouslySetInnerHTML={{ __html }}/>
  </> );
}

User-Submitted File

FileReaderSync

Typically FileReader is used in the main browser context. In Web Workers, the synchronous version FileReaderSync is more efficient.

Typically the Web Worker receives a file pointer, reads and parses the file, and sends a final result (HTML table or raw data) to the main browser context:

Live Demo (click to show)

In the following example, when a file is dropped over the DIV or when the INPUT element is used to select a file, the script:

sends the File object to the Web Worker
loads the SheetJS library and parses the file in the Worker
generates an HTML string of the first table in the Worker
sends the string to the main browser context
adds the HTML to the page in the main browser context

Result

Live Editor

function SheetJSDragDropWorker() {
  const [__html, setHTML] = React.useState("");
  /* suppress default behavior for drag and drop */
  function suppress(e) { e.stopPropagation(); e.preventDefault(); }

  /* this worker is shared between drag-drop and file input element */
  const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', (e) => {
  try {
    /* Read file data */
    const ab = new FileReaderSync().readAsArrayBuffer(e.data.file);

    /* Parse file */
    const wb = XLSX.read(ab, {dense: true});
    const ws = wb.Sheets[wb.SheetNames[0]];

    /* Generate HTML */
    const html = XLSX.utils.sheet_to_html(ws);

    /* Reply with result */
    postMessage({ html });
  } catch(e) {
    /* Pass the error message back */
    postMessage({html: String(e.message || e).bold() });
  }
}, false);
  `])));
    /* when the worker sends back the HTML, add it to the DOM */
    worker.onmessage = function(e) { setHTML(e.data.html); };
  return ( <>
    <div onDragOver={suppress} onDragEnter={suppress} onDrop={(e) => {
      suppress(e);
      /* post a message with the first File to the worker */
      worker.postMessage({ file: e.dataTransfer.files[0] });
    }}>Drag a file to this DIV to process! (or use the file input)</div>
    <input type="file" onChange={(e) => {
      suppress(e);
      /* post a message with the first File to the worker */
      worker.postMessage({ file: e.target.files[0] });
    }}/>
    <div dangerouslySetInnerHTML={{ __html }}/>
  </> );
}

function SheetJSDragDropWorker() {
  const [__html, setHTML] = React.useState("");
  /* suppress default behavior for drag and drop */
  function suppress(e) { e.stopPropagation(); e.preventDefault(); }

  /* this worker is shared between drag-drop and file input element */
  const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

/* this callback will run once the main context sends a message */
self.addEventListener('message', (e) => {
  try {
    /* Read file data */
    const ab = new FileReaderSync().readAsArrayBuffer(e.data.file);

    /* Parse file */
    const wb = XLSX.read(ab, {dense: true});
    const ws = wb.Sheets[wb.SheetNames[0]];

    /* Generate HTML */
    const html = XLSX.utils.sheet_to_html(ws);

    /* Reply with result */
    postMessage({ html });
  } catch(e) {
    /* Pass the error message back */
    postMessage({html: String(e.message || e).bold() });
  }
}, false);
  `])));
    /* when the worker sends back the HTML, add it to the DOM */
    worker.onmessage = function(e) { setHTML(e.data.html); };
  return ( <>
    <div onDragOver={suppress} onDragEnter={suppress} onDrop={(e) => {
      suppress(e);
      /* post a message with the first File to the worker */
      worker.postMessage({ file: e.dataTransfer.files[0] });
    }}>Drag a file to this DIV to process! (or use the file input)</div>
    <input type="file" onChange={(e) => {
      suppress(e);
      /* post a message with the first File to the worker */
      worker.postMessage({ file: e.target.files[0] });
    }}/>
    <div dangerouslySetInnerHTML={{ __html }}/>
  </> );
}

Streaming Write

A more general discussion, including row-oriented processing demos, is included in the "Large Datasets" demo.

XLSX.stream.to_csv incrementally generates CSV rows.

File System Access API

Browser Compatibility

At the time of writing, the File System Access API is only available in Chromium and Chromium-based browsers like Chrome and Edge.

Performance

In local testing, committing each CSV row as it is generated is significantly slower than accumulating and writing once at the end.

When the target CSV is known to be less than 500MB, it is preferable to batch. Strings larger than 500M may hit browser length limits.

Live Demo (click to show)

The following live demo fetches and parses a file in a Web Worker. The script:

prompts user to save file (window.showSaveFilePicker in the main thread)
passes the URL and the file object to the Web Worker
loads the SheetJS library in the Web Worker
fetches the requested URL and parses the workbook from the Worker
creates a Writable Stream from the file object.
uses XLSX.stream.to_csv to generate CSV rows of the first worksheet
- every 100th row, a progress message is sent back to the main thread
- at the end, a completion message is sent back to the main thread

The demo has a checkbox. If it is not checked (default), the Worker will collect each CSV row and write once at the end. If it is checked, the Worker will try to commit each row as it is generated.

The demo also has a URL input box. Feel free to change the URL. For example:

https://raw.githubusercontent.com/SheetJS/test_files/master/large_strings.xls is an XLS file over 50 MB. The generated CSV file is about 55 MB.

https://raw.githubusercontent.com/SheetJS/libreoffice_test-files/master/calc/xlsx-import/perf/8-by-300000-cells.xlsx is an XLSX file with 300000 rows (approximately 20 MB) yielding a CSV of 10 MB.

Result

Live Editor

function SheetJSFetchCSVStreamFile() {
  const [state, setState] = React.useState("");
  const [__html, setHTML] = React.useState("");
  const [cnt, setCnt] = React.useState(0);
  const [hz, setHz] = React.useState(0);
  const [url, setUrl] = React.useState("https://oss.sheetjs.com/test_files/large_strings.xlsx");
  const ref = React.useRef(null);

  return ( <>
    <b>URL: </b><input type="text" value={url} onChange={(e) => setUrl(e.target.value)} size="80"/><br/>
    <b>Commit each row: </b><input type="checkbox" ref={ref}/><br/>
    <button onClick={async() => {
      /* this mantra embeds the worker source in the function */
      const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

function sheet_to_csv_cb(ws, cb, opts, batch = 1000) {
  XLSX.stream.set_readable(() => ({
    __done: false,
    // this function will be assigned by the SheetJS stream methods
    _read: function() { this.__done = true; },
    // this function is called by the stream methods
    push: function(d) {
      if(!this.__done) cb(d);
      if(d == null) this.__done = true; },
    resume: function pump() {
      for(var i = 0; i < batch && !this.__done; ++i) this._read();
      if(!this.__done) setTimeout(pump.bind(this), 0); }
  }));
  return XLSX.stream.to_csv(ws, opts);
}

/* this callback will run once the main context sends a message */
self.addEventListener('message', async(e) => {
  try {
    /* Fetch file */
    postMessage({state: "fetching"});
    var t = Date.now();
    const res = await fetch(e.data.url);
    const ab = await res.arrayBuffer();
    postMessage({time: "fetch", ts: Date.now() - t});

    /* Parse file */
    postMessage({state: "parsing"});
    t = Date.now();
    const wb = XLSX.read(ab, {dense: true});
    const ws = wb.Sheets[wb.SheetNames[0]];
    postMessage({time: "parse", ts: Date.now() - t});

    /* Generate CSV rows */
    postMessage({state: "begin"});
    t = Date.now();
    const wstream = await e.data.wFile.createWritable();
    let c = 0, buf = "", each = !!e.data.each;
    const strm = sheet_to_csv_cb(ws, async(csv) => {
      if(csv != null) {
        if(each) await wstream.write(csv);
        else buf += csv;
        if(!(++c % 100)) postMessage({ state: "writing", c, ts: Date.now() - t });
      } else {
        if(buf) await wstream.write(buf);
        await wstream.close();
        postMessage({state: "done", c, ts: Date.now() - t });
      }
    });
    strm.resume();
  } catch(e) {
    /* Pass the error message back */
    postMessage({error: String(e.message || e) });
  }
}, false);
      `])));
      /* when the worker sends back data, add it to the DOM */
      const log = (s, t) => setHTML(h => h + `${s}: ${(t/1000).toFixed(3).padStart(8)} sec\n`);
      worker.onmessage = function(e) {
        if(e.data.error) return setState(`Processing Error: ${e.data.error}`);
        else if(e.data.state) {
          setState(e.data.state);
          if(e.data.c) setCnt(e.data.c);
          if(e.data.ts) setHz((e.data.c || cnt) * 1000 / e.data.ts);
          if(e.data.state == "done") log("write", e.data.ts);
        } else if(e.data.time) log(e.data.time, e.data.ts);
      };
      setCnt(0); setHz(0); setState(""); setHTML("");
      try {
        /* Show picker and get handle to file */
        const wFile = await window.showSaveFilePicker({
          suggestedName: "SheetJSStream.csv",
          types: [ { description: 'csv', accept: { 'text/csv': ['.csv'] } } ]
        });

        /* post a message to the worker with the URL to fetch */
        worker.postMessage({url, wFile, each: !!ref.current.checked});
      } catch(e) { setState(`Selection Error: ${e && e.message || e}`); }
    }}><b>Click to Start</b></button>
    <pre>State: <b>{state}</b><br/>Count: <b>{cnt}</b> <b>({hz|0} Hz)</b></pre>
    <pre dangerouslySetInnerHTML={{__html}}/>
  </> );
}

function SheetJSFetchCSVStreamFile() {
  const [state, setState] = React.useState("");
  const [__html, setHTML] = React.useState("");
  const [cnt, setCnt] = React.useState(0);
  const [hz, setHz] = React.useState(0);
  const [url, setUrl] = React.useState("https://oss.sheetjs.com/test_files/large_strings.xlsx");
  const ref = React.useRef(null);

  return ( <>
    <b>URL: </b><input type="text" value={url} onChange={(e) => setUrl(e.target.value)} size="80"/><br/>
    <b>Commit each row: </b><input type="checkbox" ref={ref}/><br/>
    <button onClick={async() => {
      /* this mantra embeds the worker source in the function */
      const worker = new Worker(URL.createObjectURL(new Blob([`\
/* load standalone script from CDN */
importScripts("https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js");

function sheet_to_csv_cb(ws, cb, opts, batch = 1000) {
  XLSX.stream.set_readable(() => ({
    __done: false,
    // this function will be assigned by the SheetJS stream methods
    _read: function() { this.__done = true; },
    // this function is called by the stream methods
    push: function(d) {
      if(!this.__done) cb(d);
      if(d == null) this.__done = true; },
    resume: function pump() {
      for(var i = 0; i < batch && !this.__done; ++i) this._read();
      if(!this.__done) setTimeout(pump.bind(this), 0); }
  }));
  return XLSX.stream.to_csv(ws, opts);
}

/* this callback will run once the main context sends a message */
self.addEventListener('message', async(e) => {
  try {
    /* Fetch file */
    postMessage({state: "fetching"});
    var t = Date.now();
    const res = await fetch(e.data.url);
    const ab = await res.arrayBuffer();
    postMessage({time: "fetch", ts: Date.now() - t});

    /* Parse file */
    postMessage({state: "parsing"});
    t = Date.now();
    const wb = XLSX.read(ab, {dense: true});
    const ws = wb.Sheets[wb.SheetNames[0]];
    postMessage({time: "parse", ts: Date.now() - t});

    /* Generate CSV rows */
    postMessage({state: "begin"});
    t = Date.now();
    const wstream = await e.data.wFile.createWritable();
    let c = 0, buf = "", each = !!e.data.each;
    const strm = sheet_to_csv_cb(ws, async(csv) => {
      if(csv != null) {
        if(each) await wstream.write(csv);
        else buf += csv;
        if(!(++c % 100)) postMessage({ state: "writing", c, ts: Date.now() - t });
      } else {
        if(buf) await wstream.write(buf);
        await wstream.close();
        postMessage({state: "done", c, ts: Date.now() - t });
      }
    });
    strm.resume();
  } catch(e) {
    /* Pass the error message back */
    postMessage({error: String(e.message || e) });
  }
}, false);
      `])));
      /* when the worker sends back data, add it to the DOM */
      const log = (s, t) => setHTML(h => h + `${s}: ${(t/1000).toFixed(3).padStart(8)} sec\n`);
      worker.onmessage = function(e) {
        if(e.data.error) return setState(`Processing Error: ${e.data.error}`);
        else if(e.data.state) {
          setState(e.data.state);
          if(e.data.c) setCnt(e.data.c);
          if(e.data.ts) setHz((e.data.c || cnt) * 1000 / e.data.ts);
          if(e.data.state == "done") log("write", e.data.ts);
        } else if(e.data.time) log(e.data.time, e.data.ts);
      };
      setCnt(0); setHz(0); setState(""); setHTML("");
      try {
        /* Show picker and get handle to file */
        const wFile = await window.showSaveFilePicker({
          suggestedName: "SheetJSStream.csv",
          types: [ { description: 'csv', accept: { 'text/csv': ['.csv'] } } ]
        });

        /* post a message to the worker with the URL to fetch */
        worker.postMessage({url, wFile, each: !!ref.current.checked});
      } catch(e) { setState(`Selection Error: ${e && e.message || e}`); }
    }}><b>Click to Start</b></button>
    <pre>State: <b>{state}</b><br/>Count: <b>{cnt}</b> <b>({hz|0} Hz)</b></pre>
    <pre dangerouslySetInnerHTML={{__html}}/>
  </> );
}

Installation​

Live Demos​

Downloading a Remote File​

Creating a Local File​

User-Submitted File​

Streaming Write​

File System Access API​

Installation

Live Demos

Downloading a Remote File

Creating a Local File

User-Submitted File

Streaming Write

File System Access API