Mark Beardsley <markbrdsly <at> tiscali.co.uk> writes: > > Hard to say much at all without looking at one of these files. Could you > possibly post one of the files that is causing issues for you? With your > saying that opening and then re-saving cures the problems, I suspect that > the files may contain something that is not to spec as far as POI is > concerned. Do you know how these files were created? There are report > generators that create files Excel will accept but POI will not I believe, > although I cannot name one for you at the moment. I am certain that others > know far more about this than do I but we can make a start if you could post > one of the problematical files. > > -- > View this message in context: > http://apache-poi.1045710.n5.nabble.com/Apache-POI- Streaming-API-doesn-t-recognize-Excel-xlsx-content-tp5720197p5720213.html > Sent from the POI - User mailing list archive at Nabble.com. >
I have the same issue here and can provide a sample xlsx-file. I tried to parse Excel sheets provided by this site: https://www.bondora.ee/en/invest/statistics/data_export link to file: https://www.bondora.ee/marketing/media/MonthlyOverview.xlsx The problem I found seems to be related to the namespaces used within the xmls. If the workbook.xml only contains a default namespace und uses non-prefixed tags, the content is parseable. As soon as the xml does not use that shortcut, but uses prefixed elements, the implementation doesn't work no more. I tried to fix the sample XLS2CSV (made the xml parser namespace-aware) but that didn't lead to a working version (although the parsing worked). The next issue is the parsing of the sharedStrings.xml. This one is also not namespace-aware, so I still get empty values for all cells. To me this seems like a bug in the poi library, since those Excel sheets are valid. I will look further into the issue and report back if there is a chance of implementing this without changing too much code. PS: I quick workaround is to load the excel sheet into LibreOffice and then save it again. LibreOffice than generates a xlsx-file with a default namespace. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
