Javascript to csv export encoding issue

Question

I need to export javascript array to excel file and download it I\'m doing it in this code. data is a javascript object array.

var csvContent = \"data:t         


        

Answer 1

You should add the UTF-8 BOM at the start of the text, like:

var csvContent = "data:text/csv;charset=utf-8,%EF%BB%BF";

It worked for me with Excel 2013.

Demo Fiddle

Answer 2

To export CSV containing multibyte characters and make it readable on text editor and Excel in multiple OS platforms (Windows, Linux, MacOS), the following rules should be applied:

1. Separate the field with tab instead of comma (so that Excel on MacOS can properly display the generated CSV file)
2. Encode the string / content with UTF-16 little endian (UTF16-LE) instead of UTF-8
3. Add byte order mark (BOM) 0xFEFF as specified in RFC2781 section 3.2 at the beginning of the serialized stream to explicitly provide "signature" of content encoded with UTF16-LE

Further elaboration, use cases and sample code with NodeJS can be seen in this article.

Answer 3

Excel is really bad at detecting encoding, especially Excel on OSX.

The best solution would be to encode your CSV in the default Excel encoding: windows-1252 (also called ANSI, which is basically a subset of ISO-8859-1).

I put a complete example of how to do that at: https://github.com/b4stien/js-csv-encoding.

The 2 main parts are stringencoding (to encode the content of your CSV in windows-1252) and FileSaver.js (to download the generated Blob).

It looks like:

var csvContent = 'éà; ça; 12\nà@€; çï; 13',
    textEncoder = new TextEncoder('windows-1252');


var csvContentEncoded = textEncoder.encode([csvContent]);
var blob = new Blob([csvContentEncoded], {type: 'text/csv;charset=windows-1252;'});
saveAs(blob, 'some-data.csv');

Answer 4

Option 1

use iconv-lite library and encode your output to ascii before send it back to the user. Example:

var iconv = require('iconv-lite');
buf = iconv.encode(str, 'win1255'); // return buffer with windows-1255 encoding

Option 2

Write on the head of the file the BOM header of UTF-8 encoding. Example:

res.header('Content-type', 'text/csv; charset=utf-8');
res.header('Content-disposition', 'attachment; filename=excel.csv'); 
res.write(Buffer.from('EFBBBF', 'hex')); // BOM header

// rest of your code

Option 3

Use base64 url format like data:text/csv;base64,77u/Zm9vLGJhcg0KYWFhLGJiYg==. This method will work on client-side also (IE10+, FF, Chrome, Opera, Safari).

For example:

window.location = "data:text/csv;base64,77u/" + btoa("foo,bar\r\naaa,bbb");

Answer 5

somehow found Tab-Separated-CSV with utf-16le encoding with BOM works on WIN/MAC Excel

followed b4stien's answer but make a little difference to archive:

var csvContent = 'éà; ça; 12\nà@€; çï; 13',
    textEncoder = new TextEncoder('utf-16le');
var csvContentEncoded = textEncoder.encode([csvContent]);
var bom = new Uint8Array([0xFF, 0xFE]);
var out = new Uint8Array( bom.byteLength + csvContentEncoded.byteLength );
out.set( bom , 0 );
out.set( csvContentEncoded, bom.byteLength );
var blob = new Blob([out]);
saveAs(blob, 'some-data.csv');

with Linux /usr/bin/file tests:

Little-endian UTF-16 Unicode text, with very long lines, with CRLF line terminators

Answer 6

You can add the BOM at first, use this code and try

var BOM = "\uFEFF"; 
var csvContent = BOM + csvContent;

and then crate the file headers with the data: "text/csv;charset=utf-8"