Re: problem with duplicate records

On Thu, 2008-11-20 at 16:20 +0100, nb@xxxxxxxxxxxxx wrote:

Theoretically, the refNo can be between AA00001 and ZZ99999 so only
used numbers are stored in the file. These numbers correspond
to a physical emplacement in our archives.

Won't be easier to have a second file with available number, and then
GetTheGodDamNo()simply do a setll/Reade (with lock) and delete the key ??

You cannot have have duplicate return, time will be lighting faster and you
can populated the sec file via a *AFTER *DELETE trigger so you'll be sure
every delete on the main file will result in the correct record placed back
on the second one.

As David did not raise the question of performance, I tried very hard
not to think about it. Really, I did. But I failed. Ah well, noone is
perfect.

What I came up with is basically the same solution, but packaged a bit
differently.
- A new pf duplicates the existing file plus a usage flag.
- A lf selecting active records replaces the existing file.
- A lf selecting available records gives convenient access to an
available reference number.

Cheers,
Terry.

--

Nicola Brion
Tech Fossil (Often called a Dinosaur) - ancient animal that make things to
work.

"Cinico, sm : Furfante, la cui distorta percezione della realta' lo porta a
vedere le cose come sono, e non come dovrebbero essere."