由 200 万条记录组成的 Mysql 查询组需要 7 秒

Question

I had a query with group by condition takes 7 seconds. I had optimized all its data type and also given index.

Before index Execution time: 15 sec After index and optimize data type its taken time: 7 seconds.

My database is much larger and I have to run approx 15 queries to prepare my report the single query takes 7 sec then 15*7 = 105 sec it's too much time-consuming.

Sample Query:

Select 
    ca_id, 
    count(ca_id),
    email,
    cadate
from
    tbl_zonec
group by 
    email, 
    cadate,
    ca_id 

Is there any way to optimize the query performance?

Answer 1

• Get rid of id ; it is just plain wrong not to have it in the GROUP BY . Ditto for sample .

• If ca_id cannot be NULL , say COUNT(*) .

• You seem to be scanning the entire table every time. Do you get different results? Is the table being added to? Is old data being modified or deleted, or is this effectively a write-once table? If the latter, then consider building and maintaining a Summary table ; it will be a lot faster. (Perhaps 10-fold.)

• Please provide SHOW CREATE TABLE . What is the PRIMARY KEY ? If it is does not start with email, cadate, ca_id , then there is an improvement possible.

• Without changing the PK, this would run a little faster: INDEX(email, cadate, ca_id, sample, id) -- it is optimal for the GROUP BY and it is "covering".