CONFUSING ARTICLE: Random Access Files - By Cruzan Wallis

[na_th_an]

Sure, and it's terribly simple:

Code:

GET #file%, pos&, <data>
PUT #file%, pos&, <data>

Those, respectively, get <data> from or put <data> into the file. The number of bytes depends on the length of the data type the <data> variable is. For example,

GET #1, 10, a%

Will read two bytes from position 10 and will store them in a%

PUT #1, 200, great$

Will write LEN(great$) bytes from position 200 taken from the great$ string.

You can also use them to access a file sequentially. There is a file pointer just like in INPUT and OUTPUT sequential modes and you can tell QB to read or write from that position just not telling PUT or GET a position, and that pointer gets automaticly incremented by the size of the data read or written. For example, the following code will read a binary file completely byte to byte:

Code:

byte$ = " "   ' Just to make byte$ measure 1 byte.

f%=FREEFILE
OPEN "FILE.BIN" FOR BINARY AS #f%

WHILE NOT EOF(f%)
   GET #f%, , byte$
   PRINT "Byte read: "; ASC(byte$)
WEND
CLOSE #f%

You can also write or read a full record, QB will calculate how many bytes depending on the size of the custom data type, and will increase the file pointer correctly as well:

Code:

TYPE RecordType
   id AS INTEGER
   name AS STRING * 20
   age AS INTEGER
   phone AS STRING * 10
END TYPE

DIM myRecord AS RecordType

f%=FREEFILE
OPEN "FILE.DAT" FOR BINARY AS #f%

PUT #f%, , myRecord

CLOSE #f%

or the same way around:

Code:

GET #f%, , myRecord

As you see, this is far way more flexible than the old RANDOM mode, not limiting the size of the data/write read or where to read/write that data.