what parallelism(distribute streams) operation in execution plan. why increasing rows count previous operation.
query :
select top 5000 mm.member_id, mm.member_no, mm.member_name, mm.trade_class, mt.description, ms.description, mai.address_1, mai.address_2, mai.city, mai.state, mai.zip member_master mm inner join member_type mt on mt.code = mm.member_type inner join member_status ms on ms.code = mm.member_status inner join member_address_info mai on mm.member_id = mai.member_id mai.primary_address_flag = 'yes' execution plan:

from above execution plan can see rows count increased 39,771 1,272,672. why happening?
the system has decided can benefit performing of processing in parallel. in case, it's apparently decided let 32 independent threads work set of rows (1272672 = 39771*32). shouldn't cause concern - - , "duplicate" rows eliminated later (gather streams) operation.
Comments
Post a Comment