Question

我的数据集就像这样

ID|Value
 1|10
 1|3
 2|9
 2|10
 2|15

我需要找到相同ID的两个连续元素之间的差异并存储该值，然后再次找到差异并将存储的值与具有相同ID的所有值的新差值相乘。这必须继续，直到有相同ID的元素。因此，我的最终数据集看起来应该是

ID|Value
 1|-7
 2|5

有没有人有任何建议我应该如何使用sql

Answer 1

计算连续行之间的差异 -

MariaDB [sandbox]> SELECT T.*,
    -> IF(T.TID <> @P ,@RN:=1,@RN:=@RN+1) RN,
    -> IF(T.TID <> @P ,0, @PVAL) PVAL,
    -> IF(T.TID <> @P ,0, T.VALUE * 1.00 - @PVAL) DIFF,
    -> @P:=T.TID,
    -> @PVAL:=T.VALUE
    -> FROM(SELECT @RN:=0,@P:=0,@PVAL:=0.00,@DIFF:=0.00) RN,T
    -> ORDER BY T.TID,T.dt;
+------------+------+-------+------+------+------+-----------+----------------+
| dt         | TID  | Value | RN   | PVAL | DIFF | @P:=T.TID | @PVAL:=T.VALUE |
+------------+------+-------+------+------+------+-----------+----------------+
| 2017-01-01 |    1 |    10 |    1 | 0    |    0 |         1 |             10 |
| 2017-01-02 |    1 |     3 |    2 | 10   |   -7 |         1 |              3 |
| 2017-01-01 |    2 |     9 |    1 | 0    |    0 |         2 |              9 |
| 2017-01-02 |    2 |    10 |    2 | 9    |    1 |         2 |             10 |
| 2017-01-03 |    2 |    15 |    3 | 10   |    5 |         2 |             15 |
+------------+------+-------+------+------+------+-----------+----------------+
5 rows in set (0.00 sec)

但是在sql（据我所知的任何版本）中没有聚合函数可以计算差异的产品（如果这是正确的词）。你似乎想要丢弃每个tid的第一行。你可以通过将差异group_concatenating到表中然后调用一个启动对动态sql的调用的存储过程来实现这一点（顺便说一句，你不能在存储函数中使用动态sql）。

此存储过程

drop procedure if exists f;
delimiter //

CREATE DEFINER=`root`@`localhost` procedure `F`(
    instring varchar(200)
)

LANGUAGE SQL
NOT DETERMINISTIC
CONTAINS SQL
SQL SECURITY DEFINER
COMMENT ''
begin
declare   tempstring varchar(100);
declare   outstring  varchar(100);
declare  tempid int;
declare  checkit int;
declare maxrecs int;
set checkit = 0;
set maxrecs = (select count(*) from temp);

looper: while   checkit < maxrecs do
select tid,calc into    tempid,tempstring from temp limit checkit,1; 
set checkit = checkit + 1;  
if instr(tempstring,',')  = 0 then 
    update temp
    set    rsult  = tempstring; 
else
    set outstring  = replace(tempstring,',','*');
    set outstring  = concat('update temp set rsult = (Select ',outstring,') where tid = ', tempid, ';');
    set @sqlstmt = outstring;
    #select checkit, maxrecs, outstring;
    prepare sqlstmt from @sqlstmt;  
    execute sqlstmt;
    deallocate prepare sqlstmt;
end if;

end while;
end //

delimiter ;

在此脚本中调用时会产生您想要的结果

*ariaDB [sandbox]> drop table if exists temp;
Query OK, 0 rows affected (0.09 sec)

MariaDB [sandbox]> create table temp as
    -> select s.tid,group_concat(s.diff) calc, 1.00 as rsult
    -> from
    -> (
    -> SELECT T.*,
    -> IF(T.TID <> @P ,@RN:=1,@RN:=@RN+1) RN,
    -> IF(T.TID <> @P ,0, @PVAL) PVAL,
    -> IF(T.TID <> @P ,0, T.VALUE * 1.00 - @PVAL) DIFF,
    -> @P:=T.TID,
    -> @PVAL:=T.VALUE
    -> FROM(SELECT @RN:=0,@P:=0,@PVAL:=0.00,@DIFF:=0.00) RN,T
    -> ORDER BY T.TID,T.dt
    -> ) s
    -> where s.rn <> 1
    -> group by s.tid;
Query OK, 2 rows affected (0.23 sec)
Records: 2  Duplicates: 0  Warnings: 0

MariaDB [sandbox]>
MariaDB [sandbox]>
MariaDB [sandbox]> call f('1');
Query OK, 0 rows affected (0.07 sec)

MariaDB [sandbox]>
MariaDB [sandbox]> select * from temp;
+------+-----------+-------+
| tid  | calc      | rsult |
+------+-----------+-------+
|    1 | -7.00     | -7.00 |
|    2 | 1.00,5.00 |  5.00 |
+------+-----------+-------+
2 rows in set (0.00 sec)*

获取两行之间的差异，并对差异mysql执行数学运算

1 个答案: