Daniel Morgan wrote:
>>>isn't there only a locking issue if you update the primary key of the
>>>parent table? Updating other fields should be OK. Please correct me if
>>>I'm wrong.
>>
>>and the same is true for deletes of parent-rows and inserts of
>>child-rows,...
>>the situations in which it is not necessary to have an index on a
>>foreign key are so rare in real life, it easiest not to think much and
>>just create one!
>
>
> Given that foreign keys must be to either a primary key or unique
> constraint. And given that both of these constraints build an index by
> default ... how could you not have an index on the column in the parent
> table?
>
correct me, if I'm wrong, but aren't we talking about the column in the
child table, which references the parent table?! I call this the foreign
key column; sorry if I was not precise enough;
that column in the child table does not automatically have an index,
only because it is defined as "...references father.xxx", but almost
always strongly needs one;
Received on Mon Sep 01 2003 - 10:06:30 CDT
