"Lauri Pietarinen" <[EMAIL PROTECTED]> wrote in messageWhich DBMS'es support clustering of mutiple tables except for Oracle? Is this feature really used any more?
news:[EMAIL PROTECTED]
I could now denormalise OrderDetail so that it contains cust_id also
and cluster by cust_id
(might cause you trouble down the road, if you can change the customer
of an order), in which case, with 3 I/O's I would get
- 8 customer rows
- 16 order rows
- 24 order detail rows (which would all apply to one customer)
Depending on block size, by clustering the three tables together, one might
get all of those rows for a single read potentially improving on Wol's
numbers by a factor of eight or more for this one query. Of course, doing so
would increase the cost of a table scan on the customer table.
I thought it was more trouble than worth.
regards, Lauri
---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to [EMAIL PROTECTED] so that your
message can get through to the mailing list cleanly
