分段堆叠柱状图实现
查询规定时间内的汇总堆叠柱状图效果
-
实现效果如下
1194505.png - 效果详细描述
1.起止时间可在页面设置
2.柱状的个数有限制
3.数据区分级别
时间段切分
详见“将一段时间切分为N段”https://www.jianshu.com/writer#/notebooks/41037905/notes/57957591
错误的方式
-
方式一
24段时间分别进行查询数据库，即多少个柱状执行数据库查询操作多少次，
致使，数据库资源消耗很大，下为简单的代码与sql描述
查询java
SELECT
RIGHT(DATE_FORMAT(FROM_UNIXTIME(`start`), '%Y-%m-%d %H:%i:%s'),8) TIME,
alv,
COUNT(*) COUNT FROM table_log
WHERE
START >= CAST(#{start} AS SIGNED) and
START < CAST(#{end} AS SIGNED)
GROUP BY alv
- 方式二
(SELECT RIGHT(DATE_FORMAT(FROM_UNIXTIME(`start`), '%Y-%m-%d %H:%i:%s'),8) TIME,'1463716740' AS fullstartText,alv, COUNT(*) COUNT FROM table_log WHERE START >= CAST('1463716740' AS SIGNED) AND START < CAST('1463717040' AS SIGNED) GROUP BY alv )
UNION
(SELECT RIGHT(DATE_FORMAT(FROM_UNIXTIME(`start`), '%Y-%m-%d %H:%i:%s'),8) TIME,'1463716440' AS fullstartText,alv, COUNT(*) COUNT FROM table_log WHERE START >= CAST('1463716440' AS SIGNED) AND START < CAST('1463723640' AS SIGNED) GROUP BY alv )
union一两个还很快，5六个以后就缓慢了，并且，sql语句有长度限制，超过长度需要修改数据库设置。
并且，这种方式应该只是节省了请求数据库连接的时间，sql上并未做多少改善。
解决方案
-
sql语句
sql截图
以上sql为通过程序代码拼接后的结果，执行如下
执行结果
优点是，一条sql查询出结果，缺点是，会产生一些无用的列。
- 程序调用
根据sql返回的结果进行迭代，将多条符合条件的结果捏在一起就可以了
查询"年月日时"
select
update_time as updateTime,
avg(io_total) as ioTotal
from
update_time as updateTime,
avg(io_total) as ioTotal
from
TB_DY_LUNINFO_COLL
-- 实时
where (update_time between #starttime# and #endtime#)
group by hour(update_time),minute (update_time)
order by update_time asc
-- 日
where (update_time between #starttime# and #endtime#)
group by day(update_time),hour(update_time)
order by update_time asc
-- 月
where (update_time between #starttime# and #endtime#)
group by month(update_time),day (update_time)
order by update_time asc
-- 年
where (update_time between #starttime# and #endtime#)
group by year(update_time),month (update_time)
order by update_time asc