>
>
>What kind of hardware did you do this on?

As I recall, it was a Compaq Proliant 800 (it's been a couple of years). We ran NT 4.0 Server with a very big drive array. The system had to be up 7x24. We had hardware from another company that had to access the data. We would get into it for updates, reporting, etc.

>
>I've never worked with milion+ record dbf tables personally. One developer in our org who has, claims when you get table close to a million, you get major peformance hits. Even parameterized views loading a few thousand records out of the million take forever.

This was a Fox2x application. If you use Rushmore correctly, the response is very acceptable. See my web site for info on Rushmore optimization.

>Apparently you've done ten times the size with no trouble. What do you need to do with a million record table you don't need to do with a 500,000 record table? What do you need to do with an 11 million record table you dont need to d with a million record table? I remember in the foxbase days, it would take 15 minutes to build a new index on 600,000 record table. How long does it take to build a primary key index on (say) a single integer field in an 11 million record dbf.

Indexing was the biggest problem. Everyday we would copy the table to a new DBF and index it. This was all done on the server. That way, if the index became corrupted, the backup was online and only a day old. The indexing took 4-6 hours.




>
>Thanks
>
>Gar