just how did you identify the structure of the file???
did they describe it to you or did you think it yourself.
any way (although I think you are way-off) from what I gather the file structure look like this:
the file consist of two parts a header and data parts
HEADER
----------
number_of_entries(+2) : byte (it may use integer if number of entries should > 253)
'=' : char (constant, I think it used to define the size of the previous variable)
data length (length of the path) : byte
data/path : string
date : 4 byte
I'm not sure why it adds 2 to the number of entries

DATA
------
delimiter : byte = null or 0x00
column id: byte
data length : byte
data : string
the DATA is stored sequentially so DATA1 then DATA2 then DATA3
Type
Header = record
nEntry : byte ; { byte for this data file}
equ : char ;
path_length : byte ;
path : string ;
date : Longint;
End;
Data = record
delimiter : byte;
column_id : byte;
data_length: byte;
data : string;
End;
read the header first then loop through the data (read data column by column and skip to new line when detecting column_id=1 ).
write the file should not be difficult using the same structure.