我在想, 如果表很大, 这种 join 加 group by 的写法会不会很耗资源和时间?
如果楼主手上有数据可以对比一下, 看看在 logtime 是索引的情况下, 哪种快些.

use test;
drop table if exists test;
create table test (id bigint, logtime bigint);
insert into test values (1, 5);
insert into test values (2, 6);
insert into test values (3, 6);
insert into test values (4, 8);
insert into test values (5, 8);
commit;

-- 楼上各位的写法
SELECT COUNT(1) le_count, b.logtime
FROM test,
(
SELECT DISTINCT logtime
FROM test
) b
WHERE test.logtime <= b.logtime
GROUP BY b.logtime
ORDER BY b.logtime

-- 我的写法
SELECT (
SELECT COUNT(1)
FROM test t
WHERE t.logtime <= a.logtime) as le_count, a.logtime
FROM (
SELECT DISTINCT logtime
FROM test) a;

换阿里源, 网易源, 中科大源都可以啊

一个回复都没有
看来只得另想办法了

勇气可嘉

只有我一个人觉得这个需求很奇奇怪怪吗?
"公司对于我们受理工单的时长有考核，要求全天候保证 30 分钟内将工单受理"
这个规定是在干什么啊这?

@luefei 我给的就是按每天，每个 app_id 汇总的啊
不然你觉得 group by 拿来干什么吃的呢？
可能是你被 limit0,1 迷惑了，你把 limit 去掉就看明白了

DROP TABLE IF EXISTS test.app_usage;
CREATE TABLE test.app_usage
(
log_time DATETIME,
app_id BIGINT,
log_count BIGINT
);

DELETE FROM test.app_usage;
INSERT INTO test.app_usage VALUES (NOW(), 123, 100);
INSERT INTO test.app_usage VALUES (NOW(), 456, 200);
INSERT INTO test.app_usage VALUES (NOW(), 789, 300);
INSERT INTO test.app_usage VALUES (NOW(), 124, 100);
INSERT INTO test.app_usage VALUES (NOW(), 234, 200);
COMMIT;

-- SELECT * FROM test.app_usage;
SELECT a.*
FROM (
SELECT concat(DATE_FORMAT(log_time, '%Y%m%d'), app_id) AS group_key, MAX(date_format(log_time, '%Y%m%d')) as log_date, MAX(app_id), SUM(log_count) AS daily_count
FROM test.app_usage
WHERE DATEDIFF(CURRENT_DATE(), log_time) < 4
GROUP BY group_key) AS a
ORDER BY a.daily_count DESC
LIMIT 0, 1;

-- DROP TABLE IF EXISTS test.app_usage;

现在用 gmail 转发到 hotmail 了
把关键账户认证邮箱也换到 gmail 了
其他账户不想管了，实在是太多了

@tntsec 大哥，你是不是没有理解什么叫加盐啊？