Parsing by line will load the entire line of data during each iterative loop. You need to do something with that data so an Assign is usually used within the loop. If your line of data requires additional parsing (perhaps by comma separated fields), you can nest another Data Iterator within the first to parse the fields within your line.
Parsing by length will load set bytes of data during each iterative loop. I have never used this method.
The Ignore trailing delimiter checkbox is useful if your delimited data has a null value after the last delimiter value. For example if my data looks like David,Williams,Consultant, and there isn't any data after the last comma, the Data Iterator won't try to process that null value.
My recommendation is to load your file into memory using a File Access and then to set your Data Iterator to parse the Data.