At 02:51 AM 7/08/2003 +0000, you wrote:

>Thanks for all your comments.
>
>I know this method of updating a record is not the proper way. Since
>we can just use the "Update sql statement". But as Paul said. This is
>a speed test. I just did it out of my curiosity.
>
>I will be doing the test again and made some changes like you have
>suggested and see what will be the results. By the way Daniel, I
>perform this test with the database on the same machine.
>
>Helen, do you have any suggestion on this one to improve the IBO
>performance?

Only what Jason suggests. IOW, cut out the unwanted overheads like
datasources and controls.

It's not a test I'd take seriously enough to spend time and energy advising
you about - for IBO, IBX or any other connectivity interface. Who really
cares how long it takes a client application to step through 50,000
buffered records and perform DML on each one? It's not something any
self-respecting DB developer would ever desire to do...except possibly for
a one-off data conversion from a foreign dbms, or something like that.

Helen