docs.sheetjs.com/docz/docs/09-miscellany/02-errors.md
2022-10-31 00:47:55 -04:00

9.3 KiB

sidebar_position title hide_table_of_contents
2 Troubleshooting true

Here are some common errors and their resolutions. This is not comprehensive. The issue tracker has a wealth of information and user-contributed examples.

Many of these errors have been fixed in newer releases! Ensure that the latest version of the library is being used. Some legacy endpoints are out of date. Review the Installation instructions.

If issues are not covered in the docs or the issue tracker, or if a solution is not discussed in the documentation, we would appreciate a bug report.

:::info Special Thanks

Special thanks to the early adopters and users for discovering and sharing many workarounds and solutions!

:::

Errors

"Aw Snap!" or "Oops, an error has occurred!"

Browsers have strict memory limits and large spreadsheets can exceed the limits.

Technical Limitations (click to show)

V8 (Node/Chrome) have a maximum string length that has changed over the years. Node 16 and Chrome 106 enforce a limit of 536870888 characters. This issue will manifest with error messages such as Invalid string length.

There are memory bottlenecks associated with string addresses. A number of bugs have been reported to the V8 and Chromium projects on this subject. While those bugs are being resolved, for sheets containing hundreds of thousands of rows, dense mode worksheets should be used.

The API functions support dense mode:

var wb = XLSX.read(data, {dense: true}); // creates a dense-mode sheet
XLSX.writeFile(data, "large.xlsx"); // writeFile can handle dense-mode sheets

When processing very large files is a must, consider running processes in the server with NodeJS or some other server-side technology.

If the files are small, please report to our issue tracker

SCRIPT5022: DataCloneError

IE10 does not properly support Transferable.

Object doesn't support property or method 'slice'

IE does not implement Uint8Array#slice. An implementation is included in the shim script. Check the "Standalone" Installation note

TypeError: f.substr is not a function

Some Google systems use the base64url encoding. base64url and base64 are different encodings. A simple regular expression can translate the data:

var wb = XLSX.read(b64.replace(/_/g, "/").replace(/-/g, "+"), {type:'base64'});

Error: Cannot read property '0' of undefined

FileReader#readAsText will corrupt binary data including XLSX, XLSB, XLS, and other binary spreadsheet files.

Applications should use FileReader#readAsArrayBuffer or Blob#arrayBuffer. Examples are included in "User Submissions"

Applications specifically targeting legacy browsers like IE10 should use FileReader#readAsBinaryString to read the data and call XLSX.read using the binary type.

Unsupported file undefined when reading ArrayBuffer objects

Old versions of the library did not automatically detect ArrayBuffer objects.

Workaround (click to show)

:::warning Legacy workaround

This solution is not recommended for production deployments. Native support for ArrayBuffer was added in library version 0.9.9.

:::

After reading data with FileReader#readAsArrayBuffer, manually translate to binary string and call XLSX.read with type "binary"

document.getElementById('file-object').addEventListener("change", function(e) {
  var files = e.target.files,file;
  if (!files || files.length == 0) return;
  file = files[0];
  var fileReader = new FileReader();
  fileReader.onload = function (e) {
    var filename = file.name;
    // pre-process data
    var binary = "";
    var bytes = new Uint8Array(e.target.result);
    var length = bytes.byteLength;
    for (var i = 0; i < length; i++) {
      binary += String.fromCharCode(bytes[i]);
    }
    // call 'xlsx' to read the file
    var oFile = XLSX.read(binary, {type: 'binary', cellDates:true, cellStyles:true});
  };
  fileReader.readAsArrayBuffer(file);
});

Browser is stuck!

By default, operations run in the main renderer context and block the browser from updating. Web Workers offload the hard work to separate contexts, freeing up the renderer to update.

Strange exported file names in the web browser

JS and the DOM API do not have a standard approach for creating files. There was a saveAs proposal but it was abandoned in 2014.

The library integrates a number of platform-specific techniques for different environments. In modern web browsers, the library uses the download attribute.

Third party libraries like FileSaver provide an implementation of saveAs and may help in some cases.

FileSaver.js integration (click to show)

Standalone Build

<script src="https://unpkg.com/file-saver@2.0.5/dist/FileSaver.js"></script>
<script src="https://cdn.sheetjs.com/xlsx-latest/package/dist/xlsx.full.min.js"></script>
<!-- XLSX.writeFile will use the FileSaver `saveAs` implementation -->

Frameworks and Bundlers

At the time of writing, file-saver@2.0.5 leaks saveAs to the global scope, so merely importing the module works:

import FileSaver from 'file-saver'; // as a side effect, `saveAs` is visible
import { writeFile } from 'xlsx'; // writeFile will use the global `saveAs`

Data Issues

Generated XLSX files are very large!

By default, compression is disabled. Set the option compression: true in the write or writeFile options object. For example:

XLSX.writeFile(workbook, "export.xlsx", { compression: true });

CSV and XLS files with Chinese or Japanese characters look garbled

The ESM build, used in tools like Webpack and in Deno, does not include the codepage tables by default. The "Frameworks and Bundlers" section explains how to load support.

DBF files with Chinese or Japanese characters have underscores

As mentioned in the previous answer, codepage tables must be loaded.

When reading legacy files that do not include character set metadata, the codepage option controls the codepage. Common values:

codepage Description
874 Windows Thai
932 Japanese Shift-JIS
936 Simplified Chinese GBK
950 Traditional Chinese Big5
1200 UTF-16 Little Endian
1252 Windows Latin 1

When writing files in legacy formats like DBF, the default codepage 1252 will be used. The codepage option will override the setting. Any characters missing from the character set will be replaced with underscores.

Worksheet only includes one row of data

Some third-party writer tools will not update the dimensions records in XLSX or XLS or XLSB exports. SheetJS utility functions will skip values not in range.

The following helper function will recalculate the range:

function update_sheet_range(ws) {
  var range = {s:{r:Infinity, c:Infinity},e:{r:0,c:0}};
  Object.keys(ws).filter(function(x) { return x.charAt(0) != "!"; }).map(XLSX.utils.decode_cell).forEach(function(x) {
    range.s.c = Math.min(range.s.c, x.c); range.s.r = Math.min(range.s.r, x.r);
    range.e.c = Math.max(range.e.c, x.c); range.e.r = Math.max(range.e.r, x.r);
  });
  ws['!ref'] = XLSX.utils.encode_range(range);
}
More Code Snippets (click to show)

set_sheet_range changes a sheet's range given a general target spec that can include only the start or end cell:

/* given the old range and a new range spec, produce the new range */
function change_range(old, range) {
  var oldrng = XLSX.utils.decode_range(old), newrng;
  if(typeof range == "string") {
    if(range.charAt(0) == ":") newrng = {e:XLSX.utils.decode_cell(range.substr(1))};
    else if(range.charAt(range.length - 1) == ":") newrng = {s:XLSX.utils.decode_cell(range.substr(0, range.length - 1))};
    else newrng = XLSX.utils.decode_range(range);
  } else newrng = range;
  if(newrng.s) {
    if(newrng.s.c != null) oldrng.s.c = newrng.s.c;
    if(newrng.s.r != null) oldrng.s.r = newrng.s.r;
  }
  if(newrng.e) {
    if(newrng.e.c != null) oldrng.e.c = newrng.e.c;
    if(newrng.e.r != null) oldrng.e.r = newrng.e.r;
  }

  return XLSX.utils.encode_range(oldrng);
}

/* call change_sheet and modify worksheet */
function set_sheet_range(sheet, range) {
  sheet['!ref'] = change_range(sheet['!ref'], range);
}

Adding a cell to a range

function range_add_cell(range, cell) {
  var rng = XLSX.utils.decode_range(range);
  var c = typeof cell == 'string' ? XLSX.utils.decode_cell(cell) : cell;
  if(rng.s.r > c.r) rng.s.r = c.r;
  if(rng.s.c > c.c) rng.s.c = c.c;

  if(rng.e.r < c.r) rng.e.r = c.r;
  if(rng.e.c < c.c) rng.e.c = c.c;
  return XLSX.utils.encode_range(rng);
}
range_add_cell("A1:C3","B2")

function add_to_sheet(sheet, cell) {
  sheet['!ref'] = range_add_cell(sheet['!ref'], cell);
}