Lucian,

<< I mean how can a file become corrupted and the corruption would mean
duplicating something... >>

Bytes can get shifted in the header, or someone can simply overwrite
portions of the table header with different data.

<< Stepping code only in dbisam folder, so far.  I guess it's fair to ask if
*this* changed in EDB, no? >>

Yes, EDB handles metadata completely differently.

--
Tim Young
Elevate Software
www.elevatesoft.com

Lucian,

<< After more checks, TDataFile.CheckHeader is called twice on
IndexDefs.Update. So, I now have 2 million loops. which I cut down to 1 mil,
with the suggested approach. Believe me, I can see a difference. >>

What is the actual time difference in relation to the total time ?

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim Young [Elevate Software] wrote:

> Lucian,
>
> << After more checks, TDataFile.CheckHeader is called twice on
> IndexDefs.Update. So, I now have 2 million loops. which I cut down to
> 1 mil, with the suggested approach. Believe me, I can see a
> difference. >>
>
> What is the actual time difference in relation to the total time ?

I'm having a bunch of 10 tables, so far. Only 3 do have between 400 and
500 fields. I have almost a 2.5 seconds faster response when opening
all tables. I think the 2.5 sec. comes from faster opening of those 3
tables. This is a 1.8GHz pc with IDE drives, Win2k pro.


regards,

Lucian

> > What is the actual time difference in relation to the total time ?

The total time is about 9-10 sec, initializing all 10 tables. But that
includes creating complete set of fields.

Lucian

> The total time is about 9-10 sec, initializing all 10 tables. But that
> includes creating complete set of fields.

Actually I meant the total time it takes my mainform to show up. *That*
includes opening the tables 3 times. Once to catch eventually
DBISAM_OLDVERSION and UpgradeTable, once to check fields/indexes and
restructure, and finally one more time, later, to create needed fields.

Seems I also have to optimize my scripting engine. If I could reduce
the number of times I open the table, that would be very good.

Lucian

> Seems I also have to optimize my scripting engine. If I could reduce
> the number of times I open the table, that would be very good.

Which I did. Now, for the whole thing, I am down to 3 sec with my
change in DBISAMen and ~ 4 sec without the change.

Also, I am trying a change in DBISAMtb:

procedure TDBISAMFieldDefs.Update;
begin
  DataSet.FieldDefs.Updated:=Self.Updated;
  DataSet.FieldDefs.Update;
  //Self.Updated:=DataSet.FieldDefs.Updated;//LR!
end;

I would add the third line. I have a feeling it should be there. If
this is correct it than reduces the calls to TDataFile.CheckHeader.
Any opinions, Tim?


Lucian

Lucian.

<< I would add the third line. I have a feeling it should be there. If this
is correct it than reduces the calls to TDataFile.CheckHeader.Any opinions,
Tim? >>

The whole TDBISAMFieldDefs.Update processing is particularly sensitive to
changes, so don't be surprised if it doesn't work.   On the surface,
however, it appears that it will do what you intend.

--
Tim Young
Elevate Software
www.elevatesoft.com