As well as supporting the standard files (where some line terminator [Unix: /n])
denotes the end of the line. This package can support other formats (normally length based lines).
Length based files have 2 advantages over typical unix/windows text files:
- When reading, you do not need to check every byte in the file for a <CR>. For a 600 byte line
you are probably executing 1800 instructions per line to work out the length. With a length based
file it it will be a couple of instructions,
- Ideal for storing binary data - Saves the conversion cost of converting numbers
between text & Binary (in 3rd generation languages like C and Cobol).
- Standard Unix/Windows/Mac lines.
- Fixed Record Length (ie every record is the same fixed length.
- Mainframe VB (was also used on other computers as well) - 2 Byte length at the start of the line + two byte hex zero.
- Mainframe VB Dump (sames as VB but also including the the Block Discriptor Word).
- Fujitsu Cobol VB format. 4 byte length at the start and end of the line.