I'm having trouble convincing folks but...
searching 1,000,000 times for a match on a table with 1.2million records
keymatch() 5.79 seconds, but ignores deleted records
indexseek(..,.F.) 9.1 seconds and respects deleted records
=seek() 6.78 seconds and respects deleted records.

Good old SEEK is the fastest!

keymatch is fastest but doesn't actually check the deleted status of the found record. The deleted status is a hidden field in the .dbf file. Indexes store the values of all records regardless, since if you SET DELETED OFF you would expect the index to work.


>Hi
>
>I was trying to use keymatch() to prevent duplicate records in my table but it picks up deleted records that have not yet been removed from the table. I can use indexseek() to getaround the problem but was curious as to why keymatch behaves this way???
>
>Awaiting enlightenment
>Geoff Scott