"Robinson" <toomuchspamhaspassed@myinboxtoomuchtoooften.com> wrote in
message news:ekeok9$r15$1$8302bc10@news.demon.co.uk...
>
> "Andy.I" <Andy.l@community.nospam> wrote in message
> news:28A16BB3-C348-47E9-B4D6-0DE441A72222@microsoft.com...
>> Hi
>>
>> I have a small aplication that stores data in a random access file. I'm
>> able to modify records, and add new records. But how can I delete a
>> certain record and remove it enirely from the file?
>>
>> /A.
>
> A quick and easy solution would be to "flag" it as deleted, but leave it
> in the file.  You then have another procedure that removes all deleted
> items in one go every now and then or simply overwrites a "deleted" slot
> when you add new records (i.e. the file always grows).  There is no way to
> selectively "remove" some chunk of data in the middle of a file.  You have
> to copy up to the point of deletion and append data after the point of
> deletion into a new file.  Think about a database like SQL Server, it has
> a "shrink" method and it will re-use unallocated space, but the file only
> ever gets bigger unless you actually tell it to shrink.  Deleting records
> simply returns them to the pool of usuable space.
>
> I'm not sure about the context of all of this, or whether you can modify
> your scheme significantly so it's difficult to state with any more
> certainty a better approach.
>
>