>Hi Naomi,
>Here some test results with one table with 2 005 103 records, PRG and table was palced in one folder, before each teste I quit VFP and than restart it again. I tested only SELECT, not insert and update. All tests are runned once.
>

>----------------------------------------------------------------------------------------------------------
>Buffering 5 | TableValidate 7 | FLOCK() | Table was opened | Used in SELECT      | Some records | Result
>            |                 |         | before SELECT    | WITH (BUFFERING=.t.)| was changed  |(seconds)
>----------------------------------------------------------------------------------------------------------
>   OFF             OFF            OFF          OFF                  OFF               OFF          64.286
>----------------------------------------------------------------------------------------------------------
>   OFF             ON             OFF          OFF                  OFF               OFF          45.805
>----------------------------------------------------------------------------------------------------------
>   OFF             ON             ON           ON                   OFF               OFF          82.250
>----------------------------------------------------------------------------------------------------------
>   OFF             OFF            ON           ON                   OFF               OFF          45.184
>----------------------------------------------------------------------------------------------------------
>   ON              OFF            OFF          ON                   OFF               OFF          44.371
>----------------------------------------------------------------------------------------------------------
>   ON              ON             OFF          ON                   OFF               OFF          41.110
>----------------------------------------------------------------------------------------------------------
>   ON              ON             ON           ON                   OFF               OFF          54.355
>----------------------------------------------------------------------------------------------------------
>   OFF             OFF            OFF          OFF                  ON                OFF          48.395
>----------------------------------------------------------------------------------------------------------
>   OFF             OFF            OFF          ON                   ON                OFF          49.778
>----------------------------------------------------------------------------------------------------------
>   OFF             ON             OFF          ON                   ON                OFF          71.191
>----------------------------------------------------------------------------------------------------------
>   OFF             ON             ON           ON                   ON                OFF          41.477
>----------------------------------------------------------------------------------------------------------
>   ON              OFF            OFF          ON                   ON                ON           51.246
>----------------------------------------------------------------------------------------------------------
>   ON              ON             OFF          ON                   ON                ON           57.430
>----------------------------------------------------------------------------------------------------------
>   ON              ON             ON           ON                   ON                ON           63.301
>----------------------------------------------------------------------------------------------------------
>

>
>I know that this is not a proper testing, because it doesn't have all ovvasions, and doesn't run at least twice, but I have no time for more.

Our case is the the second from bottom. Interestingly, in your case the last case with FLOCK() was slower than without.

In any case, my manager is certain that WITH BUFFERING = .T. caused this tremendeous slowdown. He proved this by reverting to the old version of the program, that loads data in ~1h. (The new version never finished).

I think we need to work with my colleague on a different algorithm. I do, of course, appreciate some input on the problem from MS or from Fabio, but our experiments proved BUFFERING = .T. to be a culprit.