On 27 Nov 2001, at 0:47, Jason Wharton wrote:

> If you set the AutoFetchAll property to true and have it load all records
> into the buffer then it will not make use of the horizontal dataset
> refinement features which require the descending index for snappy
> performance.
>
> This is highly recommended for all TIBOTable components which are going to
> be used as small lookup tables.

What I propose is to have some sort of control over horizontal
dataset refinement so that it kicks in when the user wants / is
predefine. That way the user has the option of not defining an extra
index.Right now one must define an descending index to get a
snappy performance.

(I know you have other important stuff to work on but this is a
suggestion and I think it would make the table even more smarter.)

Regards

Sandeep

Software Developer
CFL
sandeep@...
http://www.cfl.co.nz