The same (although on a fairly simple scheme, say). "Just worked." The magic of sleep mode.
I had my peace of mind because we had 100% coverage for the DAO layer. Therefore, when the schema was recreated in MS SQL, and some table and column names were updated in the mapping (donβt remember why, but the database administrators, who are alleged to be naming), we simply run our tests and do not find any unsuccessful ones .
PS I recalled one interesting detail: functional tests were all right. But when PTE started in the MS SQL database, we found that simultaneous access to one particular table was slower than on Oracle due to the spread of locks. We had to redesign this functionality.
source share