On Wed, 9 Oct 2002, Charlie Reiman wrote:
Yes, this is pretty easy. Some python script to get you going: --------------- rs = some_sql_query()
for item in rs: result_str += (str(item.DATE) + ',' + str(item.STUFF) + ',' + str(item.MORESTUFF) + '\n')
response.setHeader('Content-type', 'application/vnd.ms-excel') response.setHeader('Content-Length', len(result_str)) response.setHeader('Content-Disposition', 'attachment;filename=filename.csv') return result_str ----------------
Be warned that mozilla doesn't handle this quite right. It launches excel and hands it the data but seems to translate the EOL characters in the process so excel refuses to split the columns. Ironically, Excel 2000 only accepts the unix EOL character for csv data. So: it works fine from IE but not so fine from Mozilla. If anyone has a dual browser solution I'd like to see it. This hint wass quite helpful - at least I get reasonable effects under Linux with Galeon - this handles the resulting file with Gnumeric as default. But strangely enough it works not in every case with the very same URL returning the very same result. Sometimes Gnumeric can't handle the response and fails to interpret it. This is really strange because saving the result as file and loading it afterwards with Gnumeric works perfectly.
I've got my hands on an Excel 97 installation running under NT. Here loading of the response failed because the whole table was interpreted as a single column of a string with several comma delimited entries. I tried to create a table in this Excel version and save it as cvs. The observation was that it is not comma (',') separated but semicolon (';') separated. When I tried this the effect in Gnumeric was unchanged (working *sometimes* directly - loading after saving was always perfect). But the situation in Excel did not change at all. I just tried to write a Python script which exactly produced what I've got from Excel saving as csv - Excel failed to load it correctly. The only way to get it right into Excel was to produce a file which has each field value enclosed into '"' (no there are no ',' in any of the values), to save the file with Netscape and then Excel was able to read the file which was not possible directly. IE version 5.5 is not even able to save uch kind of file as Excel file. Resume: If there is no solid way to export robust CSV which Excel is able to cope with I have to come back to my first question: How to use Perl in Zope. (And yes, I know Excel sucks - but this does not help.) Kind regards Andreas.