0

在我的一个应用程序中，我需要计算油耗，其中考虑了完全加油和部分加油。我的 Android 应用程序基于 SQLite 语法和功能。

我尝试在 SQL 中解决这个难题，并且不想遍历所有行以编程方式完成这些事情。

结果列表中应省略 FULL='N' 的行，但应将其值添加到下一个 FULL='Y' 行。至少这是我的想法。

这是 DML（仅显示必需的列）：

CREATE TABLE FUELINGS (
ID INTEGER PRIMARY KEY,
MILEAGE INTEGER,
VOLUME DECIMAL(8, 2),
FULL CHAR(1));

这是该表的一些输入：

INSERT INTO FUELINGS VALUES(1,22995,29.48,'Y');
INSERT INTO FUELINGS VALUES(2,23385,31.27,'Y');
INSERT INTO FUELINGS VALUES(3,23869,16.71,'N');
INSERT INTO FUELINGS VALUES(4,24065,33.18,'Y');
INSERT INTO FUELINGS VALUES(5,24485,30.59,'Y');
INSERT INTO FUELINGS VALUES(6,24956,33.78,'Y');

好的，问题来了。这是我使用的查询：

SELECT T2.MILEAGE, 
       T2.VOLUME, 
       T2.FULL, 
       T2.PREVMILEAGE, 
       T2.PREVVOLUME, 
       T2.PREVFULL, 
       100 * T2.VOLUME / ( T2.MILEAGE - T2.PREVMILEAGE ) 
FROM   (SELECT T.MILEAGE, 
               T.VOLUME, 
               T.FULL, 
               T.PREVMILEAGE, 
               (SELECT VOLUME 
                FROM   FUELINGS 
                WHERE  MILEAGE = T.PREVMILEAGE 
                       AND MILEAGE IS NOT NULL 
                       AND MILEAGE > 0 
                       AND VOLUME IS NOT NULL 
                       AND VOLUME > 0) AS PREVVOLUME, 
               (SELECT FULL 
                FROM   FUELINGS 
                WHERE  MILEAGE = T.PREVMILEAGE 
                       AND MILEAGE IS NOT NULL 
                       AND MILEAGE > 0 
                       AND VOLUME IS NOT NULL 
                       AND VOLUME > 0) AS PREVFULL 
        FROM   (SELECT F.MILEAGE, 
                       F.VOLUME, 
                       F.FULL, 
                       (SELECT Max(MILEAGE) 
                        FROM   FUELINGS PREV 
                        WHERE  PREV.MILEAGE < F.MILEAGE 
                               AND MILEAGE IS NOT NULL 
                               AND MILEAGE > 0 
                               AND VOLUME IS NOT NULL 
                               AND VOLUME > 0) AS PREVMILEAGE 
                FROM   FUELINGS F 
                WHERE  MILEAGE IS NOT NULL 
                       AND MILEAGE > 0 
                       AND VOLUME IS NOT NULL 
                       AND VOLUME > 0) AS T 
        WHERE  PREVMILEAGE IS NOT NULL) AS T2 
ORDER  BY T2.MILEAGE;

这是结果：

T2.MILEAGE  T2.VOLUME   T2.FULL T2.PREVMILEAGE  T2.PREVVOLUME   T2.PREVFULL 100 * T2.VOLUME / (T2.MILEAGE - T2.PREVMILEAGE)
23385   31.27   Y   22995   29.48   Y   8.017948717948718
23869   16.71   N   23385   31.27   Y   3.4524793388429753
24065   33.18   Y   23869   16.71   N   16.928571428571427
24485   30.59   Y   24065   33.18   Y   7.283333333333333
24956   33.78   Y   24485   30.59   Y   7.171974522292993

16.92 是错误的，因为之前的加油是部分加油。我需要对这两行进行求和并计算。我知道可能有不止 1 次部分加油。但我不知道如何构建 FULL='N' 值的范围，将它们相加并将它们添加到以下 FULL='Y' 值。

可以请有人把我放在正确的方向。我怎样才能做到这一点？

提前谢谢了。

编辑/解决方案：

SELECT T.MILEAGE,
       T.PREVMILEAGE,
       SUM(VOLUME),
       100 * SUM(VOLUME) / ( T.MILEAGE - T.PREVMILEAGE )
FROM   (SELECT F1.MILEAGE      AS MILEAGE,
               MAX(F2.MILEAGE) AS PREVMILEAGE
        FROM   FUELINGS F1
               LEFT OUTER JOIN FUELINGS F2
                 ON F1.MILEAGE > F2.MILEAGE
                    AND F2.FULL = 'Y'
                    AND F2.MILEAGE IS NOT NULL
                    AND F2.MILEAGE > 0
                    AND F2.VOLUME IS NOT NULL
                    AND F2.VOLUME > 0.0
        WHERE  F1.FULL = 'Y'
               AND F1.MILEAGE IS NOT NULL
               AND F1.MILEAGE > 0
               AND F1.VOLUME IS NOT NULL
               AND F1.VOLUME > 0.0
        GROUP  BY F1.MILEAGE) AS T,
       (SELECT VOLUME
        FROM   FUELINGS F3
        WHERE  F3.MILEAGE > T.PREVMILEAGE
               AND F3.MILEAGE <= T.MILEAGE)
GROUP  BY T.MILEAGE
4

1 回答 1

0

如果您有一个名为 FuelingsFull 的表，该表仅包含有关完整加油的完整信息，则您的原始查询将起作用。这可以从Fueling中总结出来。使用窗口函数，这个过程会更简单，但你在 SQLite 中没有这些。

所以，这里有一个关于如何做到这一点的想法。首先，获取之前的完整 id：

select f1.id, max(f2.full) as prevFullId
from fuelings f1 left outer join
     fuelings f2
     on f1.id > f2.id and
        f2.full = 0
group by f1.id

现在，您拥有一组属于同一类的燃料。下一个查询将其连接回燃料，然后将它们组合在一起。

select fills.id, max(mileage) as mileage, sum(volume) as volume, count(*) as numFills
from fuelings f join
     (select f1.id, max(f2.full) as prevFullId
      from fuelings f1 left outer join
           fuelings f2
           on f1.id > f2.id and
              f2.full = 1
      group by f1.id
     ) fills
     on f.id > fills.prevFullId and f.id <= fills.id
  group by fills.id

将此作为您的 FuelingFull，您的原始查询应该可以工作。

于 2012-05-14T14:23:13.353 回答