近期在做一个小需求时，涉及一个小SQL，优化时用到了 with rollup
优化前：
SELECT
type_name typeName,
sum(num)value,
(
SELECT
sum(num)
FROM
tbl_rssmc_pro_bigcustomer_domain_type_qps
WHERE
modify_date >= '2022-05-12 10:29:52'
AND modify_date <= '2022-05-13 10:29:52'
) total
FROM
tbl_rssmc_pro_bigcustomer_domain_type_qps
WHERE
modify_date >= '2022-05-12 10:29:52'
AND modify_date <= '2022-05-13 10:29:52'
GROUP BY
type_id,
type_name
ORDER BY
sum(num) DESC
LIMIT 3
?优化后：
SELECT
*
FROM
(
SELECT
COALESCE (type_name, 'QPS总量') typeName,
sum(num) value
FROM
tbl_rssmc_pro_bigcustomer_domain_type_qps
WHERE
modify_date >= '2022-05-12 10:29:52'
AND modify_date <= '2022-05-13 10:29:52'
GROUP BY
type_name WITH ROLLUP
) a
ORDER BY
a.value DESC
LIMIT 4总结：
（1）WITH ROLLUP
???????? 用来在分组统计数据的基础上再进行统计汇总，不能与order by、limit等一起使用
（2）COALESCE (）与IFNULL() 类似
?
