Lauri Pietarinen wrote:
>Jan,
>
>> >
>> ><quote>
>> >
>> >6. D should provide some shorthand for expressing the generalized
>> >transitive closure operation, including the ability to specify
>> >generalized concatenate and aggregate operations.
>> >
>> ></quote>
>>
>> Strange, because in the prescriptions (thanks for the pointer, btw.) it
>> already says that recursion should be allowed. Any idea why then there is
>> still the need for an explicit transitive closure operation?
>>
>
>I don't necessarily have a complete answer for you but
>I could point out that the above "strong suggestion" talks
>about "providing some _shorthand_ for expressing the generalized
>transitive closure".
>
>Compare it to "Strong suggestion #2:"
>
><quote>
>
>2. D should include some declarative shorthand for expressing
>referential constraints (also known as foreign key
>constraints).
>
></quote>
Very good point, and it made me realize that I should have known the answer
myself. The answer consists probably of two parts:
- There are special efficient algorithms available to the query optimizer
for the transitive closure. And because it is not easy for the optimizer
to always recognize when a general recursive query computes a transitive
closure it is useful if the user indicates this explicitly.
- The same probabaly holds for the human reader who can now more easily
understand what it is that is computed.
Received on Sat Nov 09 2002 - 17:39:30 CST
