假设我有以下表格：
user_current_value_stats
id user_id current_total_value
1 12 175
2 14 125
3 17 170
4 18 115
value_awarded_stats_history
id user_id value_awarded date
1 12 55 2016-10-5 00:00:00+05:30
2 14 50 2016-10-5 00:00:00+05:35
3 17 70 2016-10-5 00:00:00+06:35
4 18 40 2016-10-5 00:00:00+07:34
5 12 50 2016-10-11 00:00:00+04:30
6 14 65 2016-10-11 00:00:00+04:40
7 17 75 2016-10-11 00:00:00+05:40
8 18 -35 2016-10-11 00:00:00+06:40
9 12 30 2016-10-12 00:00:00+04:30
10 14 65 2016-10-12 00:00:00+04:40
11 17 35 2016-10-12 00:00:00+05:40
12 18 65 2016-10-12 00:00:00+06:40
13 12 40 2016-10-13 00:00:00+04:40
14 14 -55 2016-10-13 00:00:00+05:40
15 17 -10 2016-10-13 00:00:00+05:45
16 18 45 2016-10-13 00:00:00+06:40
预期结果
id user_id current_total_value last_week_value difference
1 12 175 130 45
2 14 125 140 -15
3 17 170 180 -10
4 18 115 70 45
我需要
user_current_value_stats value_awarded_stats_history的用户汇总值last_week_value（将给出日期）last_week_value和current_total_value之间的差异结果应该包含以下列id，user_id，current_total_value，last_week_value，difference。
（同样current_total_value也可以作为特定用户的所有value_awarded的汇总。列value_awarded实际上是冗余数据，是value_awarded的总和来自value_awarded_stats_history的用户。）
答案 0 :(得分：1)
您必须使用子查询来计算last_week_value：
select v.user_id, sum(v.value_awarded) as last_week_value
from value_awarded_stats_history v
where v.date > (current_date - '1 week')
group by v.user_id
所以完整的查询将是这样的：
select t1.id
, t1.user_id
, t1.current_total_value
, t2.last_week_value
, (t1.current_total_value - t2.last_week_value) as difference
from user_current_value_stats t1
left outer join
(select v.user_id, sum(v.value_awarded) as last_week_value
from value_awarded_stats_history v
where v.date > (current_date - '1 week')
group by v.user_id) t2 on t2.user_id = t1.user_id