Couldn't you instead pre-process the file? For example, if you know that every line has to have a certain number of items on it, you could read the entire file content in, do a find-replace on carriage returns, and then go through the entire file content as a single item, and take every 10 items for example, and process those 10 items, then move on to the next 10, etc... until you reach the end of the file.