The number of records is not the problem - it is the layout of records, and the fields should be delimited by some character. If the record element is one character fields, parse the fields must be doable. I have always worked with mailing lists, or datasets of numbers and dates. The common delimiter is a space, but not reliable unless the records is exported that way. It helps to have a header recorder that describes each field. Without that header recorder, you must rely on visual inspection.