Primary keys and indices are essential tools for performant SQL queries, but must currently be performed manually by the user after table creation with an ALTER TABLE call (made more complex for PK's by the fact that the respective columns must be NOT NULL, which must also be defined post-facto).
At a glance, at least PK's could be relatively simply implemented by adding an argument to dbCreateTable and modifying sqlCreateTable. Indices are made more complicated by the need to define them as clustered or non-clustered, but PK's seem quite simple (says the guy who hasn't actually tried doing it).
The title is pretty self-explanatory.
Primary keys and indices are essential tools for performant SQL queries, but must currently be performed manually by the user after table creation with an
ALTER TABLE
call (made more complex for PK's by the fact that the respective columns must be NOT NULL, which must also be defined post-facto).At a glance, at least PK's could be relatively simply implemented by adding an argument to
dbCreateTable
and modifyingsqlCreateTable
. Indices are made more complicated by the need to define them as clustered or non-clustered, but PK's seem quite simple (says the guy who hasn't actually tried doing it).