POI is capable of reading but not writing some of the compressed structures that Excel uses (SharedFormula, MulBlank and MulRK records), so it is common for files to grow around 30% when POI reserializes a workbook written by Excel (The exact amount of growth depends on how many of these records are present). This issue is closely related to memory consumption - for which there has been some work done recently: https://issues.apache.org/bugzilla/show_bug.cgi?id=45570 https://issues.apache.org/bugzilla/show_bug.cgi?id=15716
So it might be the case that HSSFOptimiser has done it's job properly, but the expansion mentioned above dominates. If so, try a recent version of POI (http://encore.torchbox.com/poi-svn-build/), perhaps the recent changes will make a difference. On the other hand, if the HSSFWorkbook you are talking about was entirely created with POI (i.e. starting with the no-arg constructor), we would expect HSSFOptimiser to reduce the file size or at least, leave it unchanged. If you'd like to open a bug, please upload a sample spreadsheet and/or some java code which reproduces your problem: https://issues.apache.org/bugzilla/enter_bug.cgi On Thu, Nov 6, 2008 at 8:10 AM, Tiffany Hsu <[EMAIL PROTECTED]> wrote: > Lastly I use the new helper HSSFOptimiser.optimiseCellStyles(wb) to optimise > the HSSFWorkBook . > I found, sometimes the optimiser generated HSSFWorkBook size is bigger than > I don't use optimiser! > Does this make sense? > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
