12 KiB
title |
---|
Legacy Frameworks |
import current from '/version.js';
Over the years, many frameworks have been released. Some were popular years ago but have waned in recent years. There are still many deployments using these frameworks and it is oftentimes easier to continue maintenance than to rewrite using modern web techniques.
SheetJS libraries strive to maintain broad browser and JS engine compatibility.
Integration
The "Standalone Browser Scripts" section has
instructions for obtaining or referencing the standalone scripts. These are
designed to be referenced with <script>
tags.
Internet Explorer
:::warning
Internet Explorer is unmaintained and users should consider modern browsers. The SheetJS testing grid still includes IE and should work.
:::
The modern upload and download strategies are not available in older versions of IE, but there are approaches using older technologies like ActiveX and Flash.
Complete Example (click to show)
This demo includes all of the support files for the Flash and ActiveX methods.
- Download the standalone script and shim to a server that will host the demo:
- shim.min.js
- xlsx.full.min.js
- Download the demo ZIP to the server.
The ZIP includes the demo HTML file as well as the Downloadify support files.
Extract the contents to the same folder as the scripts from step 1
- Start a HTTP server:
npx -y http-server .
- Access the
index.html
from a machine with Internet Explorer.
Other Live Demos (click to show)
:::warning
The hosted solutions may not work in older versions of Windows. For testing, demo pages should be downloaded and hosted using a simple HTTP server.
:::
http://oss.sheetjs.com/sheetjs/ajax.html uses XMLHttpRequest to download test files and convert to CSV
https://oss.sheetjs.com/sheetjs/ demonstrates reading files with FileReader
.
Older versions of IE do not support HTML5 File API but do support Base64.
On MacOS you can get the Base64 encoding with:
$ <target_file base64 | pbcopy
On Windows XP and up you can get the Base64 encoding using certutil
:
> certutil -encode target_file target_file.b64
(note: You have to open the file and remove the header and footer lines)
Upload Strategies
IE10 and IE11 support the standard HTML5 FileReader API:
function handle_fr(e) {
var f = e.target.files[0];
var reader = new FileReader();
reader.onload = function(e) {
var wb = XLSX.read(e.target.result);
process_wb(wb); // DO SOMETHING WITH wb HERE
};
reader.readAsArrayBuffer(f);
}
input_dom_element.addEventListener('change', handle_fr, false);
Blob#arrayBuffer
is not supported in IE!
ActiveX Upload
Through the Scripting.FileSystemObject
object model, a script in the VBScript
scripting language can read from an arbitrary path on the file system. The shim
includes a special IE_LoadFile
function to read binary data from files. This
should be called from a file input onchange
event:
var input_dom_element = document.getElementById("file");
function handle_ie() {
/* get data from selected file */
var path = input_dom_element.value;
var bstr = IE_LoadFile(path);
/* read workbook */
var wb = XLSX.read(bstr, {type: 'binary'});
/* DO SOMETHING WITH workbook HERE */
}
input_dom_element.attachEvent('onchange', handle_ie);
Download Strategies
As part of the File API implementation, IE10 and IE11 provide the msSaveBlob
and msSaveOrOpenBlob
functions to save blobs to the client computer. This
approach is embedded in XLSX.writeFile
and no additional shims are necessary.
Flash-based Download
It is possible to write to the file system using a SWF file. Downloadify
implements one solution. Since a genuine click is required, there is no way to
force a download. The safest data type is Base64:
// highlight-next-line
Downloadify.create(element_id, {
/* Downloadify boilerplate */
swf: 'downloadify.swf',
downloadImage: 'download.png',
width: 100, height: 30,
transparent: false, append: false,
// highlight-start
/* Key parameters */
filename: "test.xlsx",
dataType: 'base64',
data: function() { return XLSX.write(wb, { bookType: "xlsx", type: 'base64' }); }
// highlight-end
// highlight-next-line
});
ActiveX Download
Through the Scripting.FileSystemObject
object model, a script in the VBScript
scripting language can write to an arbitrary path on the filesystem. The shim
includes a special IE_SaveFile
function to write binary strings to file. It
attempts to write to the Downloads folder or Documents folder or Desktop.
This approach can be triggered, but it requires the user to enable ActiveX. It
is embedded as a strategy in writeFile
and used only if the shim script is
included in the page and the relevant features are enabled on the target system.
Frameworks
AngularJS
AngularJS was a front-end MVC framework that was abandoned by Google in 2022. It should not be confused with "Angular" the modern framework.
The Live demo shows a simple table that is updated with file data and exported to spreadsheets.
This demo uses AngularJS 1.5.0.
Full Exposition (click to show)
Array of Objects
A common data table is often stored as an array of objects:
$scope.data = [
{ Name: "Bill Clinton", Index: 42 },
{ Name: "GeorgeW Bush", Index: 43 },
{ Name: "Barack Obama", Index: 44 },
{ Name: "Donald Trump", Index: 45 }
];
This neatly maps to a table with ng-repeat
:
<table id="sjs-table">
<tr><th>Name</th><th>Index</th></tr>
<tr ng-repeat="row in data">
<td>{{row.Name}}</td>
<td>{{row.Index}}</td>
</tr>
</table>
The $http
service can request binary data using the "arraybuffer"
response
type coupled with XLSX.read
with type "array"
:
$http({
method:'GET',
url:'https://sheetjs.com/pres.xlsx',
responseType:'arraybuffer'
}).then(function(data) {
var wb = XLSX.read(data.data, {type:"array"});
var d = XLSX.utils.sheet_to_json(wb.Sheets[wb.SheetNames[0]]);
$scope.data = d;
}, function(err) { console.log(err); });
The HTML table can be directly exported with XLSX.utils.table_to_book
:
var wb = XLSX.utils.table_to_book(document.getElementById('sjs-table'));
XLSX.writeFile(wb, "export.xlsx");
Import Directive
A general import directive is fairly straightforward:
- Define the
importSheetJs
directive in the app:
app.directive("importSheetJs", [SheetJSImportDirective]);
- Add the attribute
import-sheet-js=""
to the file input element:
<input type="file" import-sheet-js="" multiple="false" />
- Define the directive:
function SheetJSImportDirective() {
return {
scope: { opts: '=' },
link: function ($scope, $elm) {
$elm.on('change', function (changeEvent) {
var reader = new FileReader();
reader.onload = function (e) {
/* read workbook */
var ab = e.target.result;
var workbook = XLSX.read(ab);
/* DO SOMETHING WITH workbook HERE */
};
reader.readAsArrayBuffer(changeEvent.target.files[0]);
});
}
};
}
Export Service
An export can be triggered at any point! Depending on how data is represented, a workbook object can be built using the utility functions. For example, using an array of objects:
/* starting from this data */
var data = [
{ name: "Barack Obama", pres: 44 },
{ name: "Donald Trump", pres: 45 }
];
/* generate a worksheet */
var ws = XLSX.utils.json_to_sheet(data);
/* add to workbook */
var wb = XLSX.utils.book_new();
XLSX.utils.book_append_sheet(wb, ws, "Presidents");
/* write workbook and force a download */
XLSX.writeFile(wb, "sheetjs.xlsx");
Dojo Toolkit
Live Demos
The "AMD" instructions
includes details for use with require
.
Integration in the demos (click to show)
The demos use the async loading strategy with the SheetJS CDN:
<script>
dojoConfig = {
packages: [
{ name: "xlsx", location: "https://cdn.sheetjs.com/xlsx-latest/package/dist", main: "xlsx.full.min" }
]
}
</script>
<script src="//ajax.googleapis.com/ajax/libs/dojo/1.10.4/dojo/dojo.js" data-dojo-config="isDebug:1, async:1"></script>
<script>
require(["dojo/request/xhr", "xlsx"], function(xhr, _XLSX) {
/* XLSX-related operations happen in the callback */
});
</script>
The "Dojo" section in "Bundlers" includes a complete example mirroring the official example
Details (click to show)
Reading Data
When fetching spreadsheets with XHR, handleAs: "arraybuffer"
yields an
ArrayBuffer
which can be passed to XLSX.read
:
<div id="tbl"></div>
<script>
require(["dojo/request/xhr", "xlsx"], function(xhr, _XLSX) {
xhr("https://sheetjs.com/pres.numbers", {
headers: { "X-Requested-With": null },
// highlight-next-line
handleAs: "arraybuffer"
}).then(function(ab) {
/* read ArrayBuffer */
// highlight-next-line
var wb = XLSX.read(ab);
/* display first worksheet data */
var ws = wb.Sheets[wb.SheetNames[0]];
document.getElementById("tbl").innerHTML = XLSX.utils.sheet_to_html(ws);
});
});
</script>
:::note
The X-Requested-With
header setting resolves some issues related to CORS.
:::
Writing Data
XLSX.writeFile
works as expected:
<script>
require(["xlsx"], function(_XLSX) {
var ws = XLSX.utils.aoa_to_sheet(["SheetJS".split(""), [5,4,3,3,7,9,5]]);
var wb = XLSX.utils.book_new(); XLSX.utils.book_append_sheet(wb, ws, "Sheet1");
/* create an XLSX file and try to save to SheetJSDojo.xlsx */
// highlight-next-line
XLSX.writeFile(workbook, "SheetJSDojo.xlsx");
});
</script>
KnockoutJS
KnockoutJS was a popular MVVM framework.
The Live demo shows a view model that is updated with file data and exported to spreadsheets.
Full Exposition (click to show)
State
Arrays of arrays are the simplest data structure for representing worksheets.
var aoa = [
[1, 2], // A1 = 1, B1 = 2
[3, 4] // A1 = 3, B1 = 4
];
ko.observableArray
should be used to create the view model:
function ViewModel() {
/* use an array of arrays */
this.aoa = ko.observableArray([ [1,2], [3,4] ]);
}
/* create model */
var model = new ViewModel();
ko.applyBindings(model);
XLSX.utils.sheet_to_json
with header: 1
generates data for the model:
/* starting from a `wb` workbook object, pull first worksheet */
var ws = wb.Sheets[wb.SheetNames[0]];
/* convert the worksheet to an array of arrays */
var aoa = XLSX.utils.sheet_to_json(ws, {header:1});
/* update model */
model.aoa(aoa);
XLSX.utils.aoa_to_sheet
generates worksheets from the model:
var aoa = model.aoa();
var ws = XLSX.utils.aoa_to_sheet(aoa);
Data Binding
data-bind="foreach: ..."
provides a simple approach for binding to TABLE
:
<table data-bind="foreach: aoa">
<tr data-bind="foreach: $data">
<td><span data-bind="text: $data"></span></td>
</tr>
</table>
Unfortunately the nested "foreach: $data"
binding is read-only. A two-way
binding is possible using the $parent
and $index
binding context properties:
<table data-bind="foreach: aoa">
<tr data-bind="foreach: $data">
<td><input data-bind="value: $parent[$index()]" /></td>
</tr>
</table>