How to work around the [1] IE bug while saving an excel file from a Web server?
Question
I've noticed that Internet Explorer adds a number in square brackets to files downloaded from the internet (usually [1]). This creates a big problem with downloading Excel spreadsheets as square brackets are not a valid filename character inside Excel worksheet name. That problem is IE specific, others browsers are keeping same file name.
So, if you have a pivot table auto-refreshed on file opening for example, you'll get an error message saying the name "file[1].yourPivotTableName" is not valid.
Is there any solution to that problem ?
EDIT : It seems that whatever the filename suggested by HTTP directives, IE adds [1] in all cases, which cause the problem ! (So, answers about filenames aren't helpful in that case)
EDIT : I've tried some VBA code to save file under another name when it'll open. However, it doesn't work (same error message than before). Do you think there's a way to fix that with VBA ?
Solution
I've got it working using VBA provided by this cool guy (think of him fondly). It renames the file and then reattaches the pivots.
http://php.kennedydatasolutions.com/blog/2008/02/05/internet-explorer-breaks-excel-pivot-tables/
OTHER TIPS
I think that this happens when you open the spreadsheet in IE and IE saves it to a temporary file. And I think it only happens when the spreadsheet's filename has more than one dot in it. Try it with a simple "sample.xls". Another workaround is to tell users to save the file to the desktop and then open it.
It's a built-in feature in Internet Explorer.
Stop using "Open", start using "Save" in the file-download window, otherwise IE will append "[1]" to filename of the file that it places in some temporary folder.
You could build some .NET application using System.IO.FileSystemWatcher that catches the event of the creation of the downloaded file or something and renames the file.
I have solved this issue by using method where we pass 3 parameters: Filename, file extension(without the .dot) and the HTTP request); then doing the UTF-8 encoding of the filename and extension. Sample Code:
public static String encoding(String fileName, String extension, HttpServletRequest request)
{
String user = request.getHeader( "user-agent" );
boolean isInternetExplorer = ( user.indexOf( "MSIE" ) > -1 );
String var = "";
try
{
fileName = URLEncoder.encode( fileName, "UTF-8" );
fileName = fileName.trim().replaceAll( "\\+", " " );
extension = URLEncoder.encode( extension, "UTF-8" );
extension = extension.trim().replaceAll( "\\+", " " );
if ( isInternetExplorer )
{
disposition = "attachment; filename=\"" + fileName+"."+extension+"\"";
}
else
{
var = "attachment; filename*=UTF-8''" + fileName+"."+extension;
}
}
catch ( UnsupportedEncodingException ence )
{
var = "attachment; filename=\"" + fileName+"."+extension;
ence.printStackTrace();
}
return var;
}
This worked just fine in my case. Hope it will help you all.
Actually, the correct .NET-code is as following:
Response.AppendHeader("content-disposition", "attachment;filename=file.xls");
Response.ContentType = "application/vnd.ms-excel";
Note: AppendHeader, not AddHeader, which I think only works in debug web-server and IIS7.
The following has worked for me:
private string EncodeFileName(string fileName)
{
fileName = HttpUtility.UrlEncode(fileName, Encoding.UTF8).Replace("+", " ");
if (HttpContext.Current.Request.UserAgent.ToLower().Contains("msie"))
{
var res = new StringBuilder();
var chArr = fileName.ToCharArray();
for (var j = 0; j < chArr.Length; j++)
{
if (chArr[j] == '.' && j != fileName.LastIndexOf("."))
res.Append("%2E");
else
res.Append(chArr[j]);
}
fileName = res.ToString();
}
return "\"" + fileName + "\"";
}
You could just make sure that in the options box for the pivot the auto refresh is switched off. Now even when opened from the server the pivot will work perfectly
I have encountered the same problem and came up with (imo) a better solution that does not need any VBA.
If you set "Content-Disposition" header to "attachment; filename=<...>" instead of "inline; filename=<...>" the normal browsers will open dialog that will allow to save or open a file with a filename defined in a header, but Internet Explorer will behave in kind of weird way. It will open file download dialog and if you press Save it will suggest a filename that is defined in the header, but if you press Open it will save file to a temporary folder and open it with a name that is the same as your URN (without 'namespace'), e.g. if your URI is http://server/folder/file.html, so IE will save your file as file.html (no brackets, woo hoo!). This leads us to a solution:
Write a script that handles request from http://server/folder/* and when you need to serve an XLS file just redirect to that script (use your filename instead of asterisk) with Content-Disposition set to inline.
Put these four lines in your code:
response.reset();
response.setHeader("Expires", "0");
response.setHeader("Cache-Control","must-revalidate,post-check=0, pre-check=0");
response.setHeader("Pragma", "public");
Hope this helps.
In .NET I have found from experience only this seems to work for me:
Response.AddHeader("Content-Disposition", "attachment; filename=excel.xls");
Response.AddHeader("Content-Type", "application/vnd.ms-excel");
Response.ContentType = "application/vnd.ms-excel";
The duplication smells, but so far I have never got to the bottom of it (maybe Sebs post explains this). Also the "content-Disposition" value appears very finicky use a : instead of a ; or ommit the space between it and 'filename' and it blows!
Also if you have compression enabled on IIS this may fix things for you:
Response.ClearHeaders()