hive is slow?

[韶隆吴]

Hi all:
  I have two tables from hbase:orders(id bigint, total map<string,double>)
 ,orderitem(orderid bigint, productid map<string,bigint>)
and both table have more than 600000 rows.
  Now I scripts the HiveQL:select size(i.productid) from orders o join
orderitem i where o.id=i.orderid;
and it cost 426s.
 BTW.the hadoop cluster have 3 PC, I want to know where is the problem
happened.

Thanks.