Sorry for this rather noob question. I am just getting started with RapindMiner by trying it on one of the Kaggle challenge that requires categorizing user comments as insulting/not-insulting. Now there is this input CSV with tens of thousands of entries of user comments but I notice that all of rapidminer text mining operators expect document. So, am I to split each row of the CSV as a file and then feed as a set of documents to rapidminer? That doesn't seem right. What is the right way of doing it?