Which encoding opens CSV files correctly with Excel on both Mac and Windows?

We have a web app that exports CSV files containing foreign characters with UTF-8, no BOM. Both Windows and Mac users get garbage characters in Excel. I tried converting to UTF-8 with BOM; Excel/Win is fine with it, Excel/Mac shows gibberish. I'm using Excel 2003/Win, 开发者_JAVA百科Excel 2011/Mac. Here's all the encodings I tried:

Encoding  BOM      Win                            Mac
--------  ---      ----------------------------   ------------
utf-8     --       scrambled                      scrambled
utf-8     BOM      WORKS                          scrambled
utf-16    --       file not recognized            file not recognized
utf-16    BOM      file not recognized            Chinese gibberish
utf-16LE  --       file not recognized            file not recognized
utf-16LE  BOM      characters OK,                 same as Win
                   row data all in first field

The best one is UTF-16LE with BOM, but the CSV is not recognized as such. The field separator is comma, but semicolon doesn't change things.

Is there any encoding that works in both worlds?

Excel Encodings

I found the WINDOWS-1252 encoding to be the least frustrating when dealing with Excel. Since its basically Microsofts own proprietary character set, one can assume it will work on both the Mac and the Windows version of MS-Excel. Both versions at least include a corresponding "File origin" or "File encoding" selector which correctly reads the data.

Depending on your system and the tools you use, this encoding could also be named CP1252, ANSI, Windows (ANSI), MS-ANSI or just Windows, among other variations.

This encoding is a superset of ISO-8859-1 (aka LATIN1 and others), so you can fallback to ISO-8859-1 if you cannot use WINDOWS-1252 for some reason. Be advised that ISO-8859-1 is missing some characters from WINDOWS-1252 as shown here:

| Char | ANSI | Unicode | ANSI Hex | Unicode Hex | HTML entity | Unicode Name                               | Unicode Range            |
| €    | 128  | 8364    | 0x80     | U+20AC      | €      | euro sign                                  | Currency Symbols         |
| ‚    | 130  | 8218    | 0x82     | U+201A      | ‚     | single low-9 quotation mark                | General Punctuation      |
| ƒ    | 131  | 402     | 0x83     | U+0192      | ƒ      | Latin small letter f with hook             | Latin Extended-B         |
| „    | 132  | 8222    | 0x84     | U+201E      | „     | double low-9 quotation mark                | General Punctuation      |
| …    | 133  | 8230    | 0x85     | U+2026      | …    | horizontal ellipsis                        | General Punctuation      |
| †    | 134  | 8224    | 0x86     | U+2020      | †    | dagger                                     | General Punctuation      |
| ‡    | 135  | 8225    | 0x87     | U+2021      | ‡    | double dagger                              | General Punctuation      |
| ˆ    | 136  | 710     | 0x88     | U+02C6      | ˆ      | modifier letter circumflex accent          | Spacing Modifier Letters |
| ‰    | 137  | 8240    | 0x89     | U+2030      | ‰    | per mille sign                             | General Punctuation      |
| Š    | 138  | 352     | 0x8A     | U+0160      | Š    | Latin capital letter S with caron          | Latin Extended-A         |
| ‹    | 139  | 8249    | 0x8B     | U+2039      | ‹    | single left-pointing angle quotation mark  | General Punctuation      |
| Π   | 140  | 338     | 0x8C     | U+0152      | Π    | Latin capital ligature OE                  | Latin Extended-A         |
| Ž    | 142  | 381     | 0x8E     | U+017D      |             | Latin capital letter Z with caron          | Latin Extended-A         |
| ‘    | 145  | 8216    | 0x91     | U+2018      | ‘     | left single quotation mark                 | General Punctuation      |
| ’    | 146  | 8217    | 0x92     | U+2019      | ’     | right single quotation mark                | General Punctuation      |
| “    | 147  | 8220    | 0x93     | U+201C      | “     | left double quotation mark                 | General Punctuation      |
| ”    | 148  | 8221    | 0x94     | U+201D      | ”     | right double quotation mark                | General Punctuation      |
| •    | 149  | 8226    | 0x95     | U+2022      | •      | bullet                                     | General Punctuation      |
| –    | 150  | 8211    | 0x96     | U+2013      | –     | en dash                                    | General Punctuation      |
| —    | 151  | 8212    | 0x97     | U+2014      | —     | em dash                                    | General Punctuation      |
| ˜    | 152  | 732     | 0x98     | U+02DC      | ˜     | small tilde                                | Spacing Modifier Letters |
| ™    | 153  | 8482    | 0x99     | U+2122      | ™     | trade mark sign                            | Letterlike Symbols       |
| š    | 154  | 353     | 0x9A     | U+0161      | š    | Latin small letter s with caron            | Latin Extended-A         |
| ›    | 155  | 8250    | 0x9B     | U+203A      | ›    | single right-pointing angle quotation mark | General Punctuation      |
| œ    | 156  | 339     | 0x9C     | U+0153      | œ     | Latin small ligature oe                    | Latin Extended-A         |
| ž    | 158  | 382     | 0x9E     | U+017E      |             | Latin small letter z with caron            | Latin Extended-A         |
| Ÿ    | 159  | 376     | 0x9F     | U+0178      | Ÿ      | Latin capital letter Y with diaeresis      | Latin Extended-A         |

Note that the euro sign is missing. This table can be found at Alan Wood.

Conversion

Conversion is done differently in every tool and language. However, suppose you have a file query_result.csv which you know is UTF-8 encoded. Convert it to WINDOWS-1252 using iconv:

iconv -f UTF-8 -t WINDOWS-1252 query_result.csv > query_result-win.csv

For UTF-16LE with BOM if you use tab characters as your delimiters instead of commas Excel will recognise the fields. The reason it works is that Excel actually ends up using its Unicode *.txt parser.

Caveat: If the file is edited in Excel and saved, it will be saved as tab-delimited ASCII. The problem now is that when you re-open the file Excel assumes it's real CSV (with commas), sees that it's not Unicode, so parses it as comma-delimited - and hence will make a hash of it!

Update: The above caveat doesn't appear to be happening for me today in Excel 2010 (Windows) at least, although there does appear to be a difference in saving behaviour if:

• you edit and quit Excel (tries to save as 'Unicode *.txt')

compared to:

• editing and closing just the file (works as expected).

The lowdown is: There is no solution. Excel 2011/Mac cannot correctly interpret a CSV file containing umlauts and diacritical marks no matter what encoding or hoop jumping you do. I'd be glad to hear someone tell me different!

You only have tried comma-separated and semicolon-separated CSV. If you had tried tab-separated CSV (also called TSV) you would have found the answer:

UTF-16LE with BOM (byte order mark), tab-separated

But: In a comment you mention that TSV is not an option for you (I haven't been able to find this requirement in your question though). That's a pity. It often means that you allow manual editing of TSV files, which probably is not a good idea. Visual checking of TSV files is not a problem. Furthermore editors can be set to display a special character to mark tabs.

And yes, I tried this out on Windows and Mac.

Here's the clincher on importing utf8-encoded CSV into Excel 2011 for Mac: Microsoft says: "Excel for Mac does not currently support UTF-8." Excel for Mac 2011 and UTF-8

Yay, way to go MS!

The best workaround for reading CSV files with UTF-8 on Mac is to convert them into XLSX format. I have found a script made by Konrad Foerstner, which I have improved little bit by adding support for different delimiter characters.

Download the script from Github https://github.com/brablc/clit/blob/master/csv2xlsx.py. In order to run it you will need to install a python module openpyxl for Excel file manipulation: sudo easy_install openpyxl.

In my case this worked (Mac, Excel 2011, both Cyrillic and Latin characters with Czech diacritics):

• Charset UTF-16LE (simply UTF-16 was not enough)
• BOM "\xFF\xFE"
• \t (tab) as separator
• Don't forget to encode also separator and CRLFs :-)
• Use iconv instead of mb_convert_encoding

It seems to my case that Excel 2011 for Mac OS is not using Encoding.GetEncoding("10000") as i thought and wasted 2 days with but the same iso as on Microsoft OS. The best proof for this is to make a file in Excel 2011 for MAC with special chars, save it as CSV and then open it in MAC text editor and the chars are scrambled.

For me this approach worked - meaning that csv export on Excel 2011 on MAC OS has special western europeean chars inside:

Encoding isoMacOS = Encoding.GetEncoding("iso-8859-1");
Encoding defaultEncoding = Encoding.Default; 

// Convert the string into a byte array.
byte[] defaultEncodingBytes = defaultEncoding.GetBytes(exportText);

// Perform the conversion from one encoding to the other.
byte[] ansiBytes = Encoding.Convert(defaultEncoding, isoMacOS, defaultEncodingBytes);

decodedString = isoMacOS.GetString(ansiBytes);

UTF-8 with no BOM currently works for me in Excel Mac 2011 14.3.2.

UTF-8 + BOM kind of works, but BOM rendered as gibberish.

UTF-16 works if you Import the file and complete the wizard, but not if you just double-click it.

The following worked for me on Excel for Mac 2011 and Windows Excel 2002:

1. Using iconv on Mac, convert the file to UTF-16 Little-Endian + name it *.txt (the .txt extension forces Excel to run the Text Import Wizard):

   iconv -f UTF-8 -t UTF-16LE filename.csv >filename_UTF-16LE.csv.txt

2. Open the file in Excel and in the Text Import Wizard choose:

   • Step 1: File origin: ignore it, it doesn't matter what you choose
   • Step 2: select proper values for Delimiters and Text qualifier
   • Step 3: if necessary, select column formats

PS The UTF-16LE created by iconv has BOM bytes FF FE in the beginning.

PPS My original csv file was created on a Windows 7 computer, in UTF-8 format (with the BOM bytes EF BB BF in the beginning) and used CRLF line breaks. Comma was used as field delimiter and single quote as text qualifier. It contained ASCII letters plus different latin letters with tildes, umlaut etc, plus some cyrillic. All displayed properly in both Excel for Win and Mac.

PPPS Exact software versions:
* Mac OS X 10.6.8
* Excel for Mac 2011 v.14.1.3
* Windows Server 2003 SP2
* Windows Excel 2002 v.10.2701.2625

On my Mac OS, Text Wrangler identified a CSV file created with Excel as having "Western" encoding.

After some googling I have made this small script (I am not sure about Windows availability, maybe with Cygwin?):

$ cat /usr/local/bin/utf8.sh

#!/bin/bash

INPUTFILE="$1"

iconv -f macroman -c -t UTF-8 $INPUTFILE |tr '\r' '\n' >/tmp/file.$$.csv

mv $INPUTFILE ms_trash
mv /tmp/file.$$.csv $INPUTFILE

In my case adding Preamble to file solved my problem:

var data = Encoding.UTF8.GetBytes(csv);
var result = Encoding.UTF8.GetPreamble().Concat(data).ToArray();
return File(new MemoryStream(result), "application/octet-stream", "data.csv");

instead of csv, trying outputting html with an XLS extension and "application/excel" mime-type. I know this will work in Windows, but can't speak for MacOS

This works for me

1. Open the file in BBEdit or TextWrangler*.
2. Set the file as Unicode (UTF-16 Little-Endian) (Line Endings can be Unix or Windows). Save!
3. In Excel: Data > Get External Data > Import Text File...

Now the key point, choose MacIntosh as File Origin (it should be the first choice).

This is using Excel 2011 (version 14.4.2)

*There's a little dropdown at the bottom of the window

Solve this using java ( UTF-16LE with BOM ):

String csvReportStr = getCsvReport();
byte[] data = Charset.forName("UTF-16LE").encode(csvReportStr)
    .put(0, (byte) 0xFF)
    .put(1, (byte) 0xFE)
    .array();

Note that CSV file should use TAB as separator. You can read the CSV file both on windows and MAC OS X.

Refer to: How do I encode/decode UTF-16LE byte arrays with a BOM?