计算mysql中行之间的差异之和

Question

How do I calculate the difference between each row and return the total sum and also how to write it in a procedure.(I use )

(There will be gaps in the id column)

id --- session --- value
1.........3..........15
4.........3...........5
7.........3...........8
10........3..........13
11........3...........9

From id.1 to id.4 difference is
From id.2 to id.3 difference is
From id.3 to id.4 difference is
From id.4 to id.5 difference is
The total sum is =

This is what I have come up with: (and yes its completely wrong)

SET @oldV = 0;

SELECT SUM(table.x)
FROM(
    SELECT @oldV := value AND IF(value > @oldV, @x:=(value - @oldV), @x:=(@oldV - value)) x
    FROM TheTable WHERE session = 3 ORDER BY id) table;

Answer 1

based on your example you seem to be calculating the absolute difference between rows. You can do this with a join statement. Try something like the following:

select a.id,b.id,abs(a.value - b.value) as dif
from UserData as a 
inner join
UserData as b
on a.id = b.id - 1

then once you have the dif col you can do a sum. If there are going to be gaps in the id create a new col ROWNUMBER which has no gaps and then use that instead of id as your on statement

Answer 2

http://sqlfiddle.com/#!9/c4573/5

SELECT *, 
   IF(@prev IS NULL,0, @diff := ABS(`value`-@prev)),
      @prev:=`value`,
      @total:=IF(@total IS NULL,0,@total)+@diff
FROM table1
ORDER BY id

Answer 3

There's lots of ways to obtain the specified result.

One possibility (specific to MySQL) is to use a user-defined variable to store the "value" from the previous row, so you can compare that to the "value" from the current row.

Note: the behavior of MySQL user-defined variables in a statement like the one below is undefined. The MySQL Reference Manual warns against depending on the behavior of user-defined variables in a context like this. BUT... we do observe repeatable consistent behavior in MySQL 5.1 and 5.5.

SELECT SUM(d.diff) AS tot_diff
  FROM ( SELECT ABS(t.`value`-@pvalue) AS diff
              , @pvalue := t.`value`
           FROM mytable t
          CROSS
           JOIN ( SELECT @pvalue := NULL ) i
          ORDER BY t.id
       ) d

Note that it's important that the user-defined variable be initialized, since the variable persists for the session, not just the statement. That's why we include the initialization in an inline view (derived table i ). Optionally, we could run a separate SET @pvalue := NULL; statement before we run the query.

Answer 4

I do not like it very much, but it seems to work.

SELECT 
   SUM(
       ABS(t1.num - t2.num)
   ) AS sumofdiffs
   FROM sumdiff AS t1 
   JOIN sumdiff AS t2 
      ON(
          (SELECT MIN(id) FROM sumdiff AS t3 WHERE t3.id > t1.id) = t2.id
      )
;

Found another way. Using just a subquery, no joins. It still is rather slow. But should be better thant the first query.

SELECT 
    SUM( 
        ABS( t1.num - 
            (SELECT num FROM sumdiff WHERE id > t1.id ORDER BY id ASC LIMIT 1 ) 
        )
    ) AS summofdiffs
FROM sumdiff AS t1;

SQLFiddle