Data with special character getting ignored while training

Hello,

I have a training model and i train that using csv data set. 
I came across one particular observation where if the row in csv data set contains special characters like \n,\r ot ^M (which can come easily if files are generated on windows and transfered to linux.) , the training process just ignores all these lines.
My csv data contains these specical characters on almost every line and hence the model is not getting trained with all these data as is ignoring these rows with special characters.

What is the way to instruct rapidminer server to allow these characters.

Please Note: I dont have control over the special characters coming in the CSV. I can do a lot of regular expression replacement in the csv to remove those special characters  but that is the last option.