Created attachment 31217 [details] diff for this patch With this patch I've aimed to output text from the XSSFEventBasedExcelExtractor in the same order as XSSFExcelExtractor, namely: - sheet name - first header - odd header - even header - cell text - textbox text - first footer - odd footer - even footer The downside of this change is that the processing can't be done as elegantly as before using a single StringBuffer. Now the SheetTextExtractor has to collect the cell text in its own StringBuffer and append it to the output to accommodate the fact that the headers and footers are processed only after the rows and cells.
Created attachment 31218 [details] test file to go in test-data/spreadsheet
Thanks for this, applied with minor changes in r1563657.