我正在尝试运行一个按日期字段对联系人进行分组的查询,然后返回每个分组的计数。如果日期字段具有作为一个组的空值,如果日期字段是10年以前是另一个组,如果日期字段是3 - 10年前,那是第三个分组,并且如果日期字段是< 3年前那是最后的分组。
对此的任何帮助将不胜感激。
答案 0 :(得分:0)
您需要在GROUP子句中使用CASE语句:
SELECT CASE
WHEN floor(DATEDIFF(EndDate, StartDate)/365) < 3 THEN '<3'
WHEN floor(DATEDIFF(EndDate, StartDate)/365) < 10 THEN '3-10'
WHEN floor(DATEDIFF(EndDate, StartDate)/365) = 10 THEN '=10'
ELSE '>10'
END AS `Length`
FROM ...
WHERE ...
GROUP BY CASE
WHEN floor(DATEDIFF(EndDate, StartDate)/365) < 3 THEN '<3'
WHEN floor(DATEDIFF(EndDate, StartDate)/365) < 10 THEN '3-10'
WHEN floor(DATEDIFF(EndDate, StartDate)/365) = 10 THEN '=10'
ELSE '>10'
END
ORDER BY Length ASC;
如果您的结束日期是现在,请使用NOW()
。
答案 1 :(得分:0)
http://sqlfiddle.com/#!9/86b630/2
SELECT IF(my_date IS NULL, 'NO DATE',
IF(YEAR(NOW())-YEAR(my_date)>10,'>10',
IF(YEAR(NOW())-YEAR(my_date)>3,'3-10','<3')
)
) `period`, COUNT(id)
FROM contacts
GROUP BY `period`
答案 2 :(得分:0)
您可以使用子选择来准备值,然后解析它们并将它们分组:
select case
when elapsed is null then 'missing'
when elapsed <= 1095 then 'new'
when elapsed <= 3650 then 'middle'
else 'old'
end as period, count(*)
from
(select datediff(now(), yourdate) period from yourtable) t
group by period