如何提高SQL查询性能

Question

I Have a view that return to me the customers and sum of their payment and its work but it take 8 second and when i remove the computed column part from the query it didn't take any time, and i tried to solve this by creating a non-clustered index on tables but its also 8 second, so I'm asking about if there is any method to improve the performance, although i have to use this technique.

my code is

SELECT ID
    ,NAME
    ,Total = (
        SELECT Sum(Value)
        FROM TRANSACTION
        WHERE Employee.ID = employee_ID
        )
FROM Employee;

Thanks in advance

Answer 1

You can try to use a join with group and see if that is faster:

 select employee.id, employee.name, sum(value)
 from employee 
 join transaction where transaction.employee_ID = employee.id
 group by employee.id, employee.name

Also - try to avoid to give names that correspond to reserved words - like TRANSACTION - see https://dev.mysql.com/doc/refman/5.5/en/keywords.html

To hasten up you might want to consider a combined index on Emlpoyee(id,name) AND an index on Transaction.Employee_ID - both are neede to make the joins and grouping for this query fast. Check if transaction.employee_ID has a FK_Constraint to Employee.ID - that might help as well.

Answer 2

You could use a join in the subselect with the sum

select e.id, e.name, t.total 
from employee  e
join (
  select employee_ID, sum(value) as total
  from TRANSACTION
  group by employee_ID
) t on t.employee_ID = e.id 

Answer 3

For this query:

SELECT ID, NAME,
       (SELECT Sum(t.Value)
        FROM TRANSACTION t
        WHERE e.Employee.ID = t.employee_ID
       ) as Total
FROM Employee e;

You want an index on TRANSACTION(employee_ID, Value) . That should actually give you optimal performance.

You can write this as a join -- but a left join :

SELECT e.ID, e.NAME, SUM(t.Value) as Total
FROM Employee e JOIN
     TRANSACTION t
     ON e.Employee.ID = t.employee_ID
GROUP BY e.ID, e.NAME;

However, the overall GROUP BY is usually less efficient than the "sub group by" in MySQL (and often in other databases as well).

Answer 4

Try an indexed view on the total

CREATE VIEW totalView
WITH SCHEMABINDING
AS
    select employee_ID, sum(value) as total
      from TRANSACTION
      group by employee_ID

CREATE UNIQUE CLUSTERED INDEX PK_totalView ON totalView
(
    employee_id,total
);

Then join to that

select employee.id, employee.name, coalesce(t.total,0)
 from employee 
 left outer join totalView t where t.employee_ID = employee.id

Adjust the join type according to your need.