MySql如何获取相邻数据

如何获取相邻数据

因为项目，所以找到了一些资料并且总结了下关于获取相邻数据的方式。

我只找到了以下的...

/*获取id值与5相减绝对值最近的数据,如果有绝对值相同的,相邻的数据优先输出*/
/*ABS(X)其实就是获取绝对值，然后的话order by 后面的1代表着根据select关键字
后的第一个字段进行排序。limit 后面的数字代表着你想查询数据的行数*/
SELECT ABS(cid-5),cname FROM company ORDER BY 1 LIMIT 3

/*获取指定id的上一条记录，我这里是以id=5为例的*/
select cid,cname from company order by cid>=5,cid desc limit 1

/*获取指定id的下一条记录，我这里是以id=5为例的*/
SELECT cid,cname from company order by cid<=5,cid asc limit 1

记录一下，以免以后遇到又不会

同表相邻数据比对查询

需求

我们将会比对相邻的数据，其中value是递增的，但也会存在清零的情况。我们的需求是计算当天的分钟递增量。

SQL

语句中需要解释一下参数

** ctc_etl.1_1_1907/sum.out_2021 ：表名**** WHERE
item_timestamp > “2021-12-01 00:00:00”
AND item_timestamp < “2021-12-02 00:00:00” 是我自己加的筛选条件**

SELECT b.id,( b.item_value - a.item_value ) AS value,
b.item_timestamp AS time
FROM
(
SELECT id,item_value,
@num := @num + 1 AS row_number
FROM
( SELECT @num := 0 ) r,
ctc_etl.`1_1_1907/sum.out_2021`
WHERE
item_timestamp > "2021-12-01 00:00:00"
AND item_timestamp < "2021-12-02 00:00:00"
ORDER BY
id
) a,
(
SELECT
id,
item_value,
item_timestamp,
@num2 := @num2 + 1 AS row_number
FROM
( SELECT @num2 := 0 ) r2,
ctc_etl.`1_1_1907/sum.out_2021`
WHERE
item_timestamp > "2021-12-01 00:00:00"
AND item_timestamp < "2021-12-02 00:00:00"
ORDER BY
id
) b
WHERE
a.row_number + 1 = b.row_number;

解析

目前计算的只是每秒的增加量，其次就是存在负数的情况，也就是清零了后一个数比前一个小就造成了负数

最终SQL

SELECT
c.id,
SUM( c.VALUE ) AS value,
DATE_FORMAT( c.time, "％Y-％m-％d ％H:％i:00" ) AS time
FROM
(
SELECT
b.id,
(b.item_value - a.item_value ) AS value,
b.item_timestamp AS time
FROM
(
SELECT
id,
item_value,
@num := @num + 1 AS row_number
FROM
( SELECT @num := 0 ) r,
ctc_etl.`1_1_1907/sum.out_2021`
WHERE
item_timestamp > "2021-12-01 00:00:00"
AND item_timestamp < "2021-12-02 00:00:00"
ORDER BY
id
) a,
(
SELECT
id,
item_value,
item_timestamp,
@num2 := @num2 + 1 AS row_number
FROM
( SELECT @num2 := 0 ) r2,
ctc_etl.`1_1_1907/sum.out_2021`
WHERE
item_timestamp > "2021-12-01 00:00:00"
AND item_timestamp < "2021-12-02 00:00:00" ORDER BY id ) b WHERE a.row_number + 1 = b.row_number ) c WHERE c.VALUE > 0
GROUP BY
DATE_FORMAT(
c.time,
"％Y-％m-％d ％H:％i:00")

来源：https://blog.csdn.net/Rude_M/article/details/81133681