Re: Optimize Query for ndb 7.2.1....

[Massimo Brignoli]

Hi Munazir,

how many rows are retrieved by the query *before* the group by?

if you execute

Select count(*)
FROM cbp_pcetcontract A,
cbp_bpprofile B,
cbp_position C,
cbp_contractdeliverable D,
cbp_contract E
where A.BP_Id = B.BP_Id
and A.BP_Id = B.BP_Id
and A.PositionId = C.Position_ID
and D.Contract_ID = E.Contract_ID
and A.ContractDeliverableId=D.CD_Id

what is the result?

Remember that the joins are pushed down but the group by and order by are still executed in the MySQL server, not in the data nodes.
So if the count(*) gives 100.000 rows for example, the data nodes in 7.2 execute the joins but they return 100.000 rows to the mysql server through the network and then the mysql server group and order them.

Massimo