Hi, When I open a .xlsx file in POI is see a lot of instance of class org.apache.xmlbeans.impl.store.Xobj. They take up a large space in the JVM( compare to the .xls file ). Is it a bug in the memory usage when parsing OOXML files? Or it is just a different in the implementation when parsing old and new excel files?
The XSSF UserModel needs more memory than the equivalent HSSF one, because it's XML based rather than records, and that uses more. If memory is a real issue, you'll want to do SAX based processing. http://poi.apache.org/spreadsheet/how-to.html#xssf_sax_api details how