Mikito Harakiri wrote:
>hidders_at_hcoss.uia.ac.be (Jan Hidders) wrote in message news:<3dd12e86$1_at_news.uia.ac.be>...
>> > But how on earth does the optimiser handle this
>> >query besides materialising the whole tree?
>>
>> By pushing the selection into the recursion and using a query optimization
>> technique called "magic sets".
>
>Consider a simple recursion example:
>
>with iota (n) as
>( values(1)
> union all
> select n+1 from iota
> where n<100000
>)
>select * from iota where n*n=25;
>
>Could n*n=25 predicate be pushed into the recursion? When a predicate
>can be pushed into recursion?
That's not so easy to explain. If you have "the Alice book" see section 13.3
on ... Magic :-).
Received on Thu Nov 14 2002 - 03:32:46 CST
