hello all,

i have an interesting question for all of you reg table/query performance.

i have one table with a clustered composite primary key and 2 Foreign keys which i am trying to modify in a stored proc. my Delete takes couple of minutes on this table.

for experiment, i have created/cloned the above table with only one exception. this new table has no Foreign keys.

now in a batch, i drop the 2 Foreign keys from my original table. i run the stored proc and then i recreate the 2 Foreign keys.

now here is the interesting puzzle.
in the stored proc, right after the orig table Delete, i also Delete the new table in the same fashion. guess what?

the orig table takes the same amount of time (even after dropping the FKs) whereas the new table takes less than a second.

can somebody explain me why do not i see the same time for Delete on my original table? (that is, despite at the time of Delete, its structure is the same as the newly cloned table.)

i appreciate your help.
many thanks.