Hi Greg,
The ability to control moving the record pointer is indeed the functional difference between INDEXSEEK and SEEK. But it's also not necessary to call it twice. The key here is when looking for potential optimizations, I substituted the equivalent INDEXSEEK() for SEEK() in the GetCellRecord method and it was ~4-5% faster with my test dataset. This implies that the underlying VFP core code for INDEXSEEK has been optimized a bit more than SEEK.
Having said that, I have not yet seen similar performance bumps for the other places where SEEK is being used but I also haven't generated coverage logs for this stuff.
↧