Mason_S
Senior member
Joined: 1 Apr 2008 09:03:08
Messages: 380
Offline
|
Hi Wu,
It depends on the particular query. If you execute a SELECT count(*) for example, it will execute a SELECT count(*) on each of the nodes, get back a one row result, then sum those up for the final result.
If doing a query where you join on two tables where the join columns are not partitioned columns, it will end up shipping around a lot of data to the nodes for joining. Ideally, you should try and construct your schema to avoid this if possible. (We will soon commit changes that speed up this handling.)
Regards,
Mason
|