Standard Random File logic employs the use of pointers embedded either in the
records or in an external (second) random file, and you "delete" records
logically by removing the pointers from the "active" chain and attaching them
to the "delete" chain.

Any "add" operation would then check the "delete" chain first for an available
record slot before grabbing more disk space.

To physically remove records from such a file, you'd have to copy all the
"active" records to a new file, erase the old file, and rename the new file.

Note that this approach can be used more effectively with BINARY files than
with RANDOM files (but why I say that is too involved for me to go into at
this point).

Hope that proves to be of some help.

--
  ++   ++   "Well Samwise: What do you think of the elves now?"


  ++   ++------------------------------------------------------

1)      Use ISAM instead - It does it for you.
2)      Copy the records (Less The blank) to another file. Then delete
        the original and rename the new file back to the old name.
3)      Move the blank record to the end of the file and then using
        a library routine written by QuickPac Pro (From
        Crescent Software) Clip the file by the length of the record.

OR -

4)      Move the blanks to the end of the file and then ignore them.
        when ever you say eof(1) say eof(1)- NoOfBlanks.
        Off course when you add another record make sure it goes into
        a blank record.

Hope this helps

Phil