>Hi Naomi,
>>Do you know if filtered indexes allow optimization in VFP9 SP1?
>
>AFAIR a filtered index ( FOR...) or one with "not" in its expresion
>is not rushmore-usable.
>
>Most of the time a
single compound index is faster for querying or joining
>(iif(iActive_flag, "1", "0") + cTrans_fk)
>but "overindexing" slows down insert/update operation. Since changes in PK/FK are
>probably not happening very often, you'ld have to decide if the speed gain
>from using only one compound rushmore bitmap vs. 2 single bitmaps or
>vs. one single bitmap and record checks on the logical field is worth
>the added overhead writing to the cdx.
>
>regds
>
>thomas
The field is Integer (iActive_Flag <g>), but I think there is a change in VFP9 regarding filtered indexes. I'm wondering, if it's documented in Help (it's too late right now to look for VFP Focus CoDe magazine issue, where I read it).
If it's not broken, fix it until it is.
My Blog