Please note that I am talking about .doc and not .docx. Parsing the same table in .docx works like a charm (but I can not convert/upgrade all files to .docx). The above is the best I could come out with to extract tables from .doc. It just misses a clear identification of end-of-row. But since I can easily spot this end-of-row if I know the number of columns, there must be a way to automate this. All my attempts with regex failed though.