A system and method for automatically determining the file record format and characteristics for a data file of unknown file record format. At least a portion of an unknown file is obtained from a file source, such as a mainframe computer. At least a portion of the file is examined to determine whether the file is EBCDIC or ASCII encoded. File type, including print, card, text or binary, is determined. File headers and trailers are detected and removed. Line delimiters such as carriage return, linefeed or new line characters are detected. Periodic character patterns are searched for, such as a space followed by one or more numeric digits, to determine file record length and contents. Once the file record format is determined, a user may verify and modify the data. This system and method can be used by a data processing system, including a telephony system which accesses data records to determine parties in a database to contact, and automatically dial those parties.