使用带数据透视的 SQL 将数据拆分为每年

Question

我在弄清楚如何构建一系列 SQL 语句以将我拥有的一些数据拆分为每年的数据时遇到了一些麻烦。在这么多年中，有 n 个项目在一组天的记录数量。这是我拥有的主要数据的一个小示例布局。

   |========|==============|===========|
   | Name   | Date         | Quantity  |
   |========|==============|===========|
 1 | AAAAA  | 10-DEC-2008  | 5         |
 2 | AAAAA  | 11-DEC-2008  | 2         |
 3 | AAAAA  | 12-DEC-2008  | 0         |
 4 | AAAAA  | 09-DEC-2009  | 3         |
 5 | AAAAA  | 10-DEC-2009  | 2         |
 6 | AAAAA  | 11-DEC-2009  | 3         |
 7 | BBBBB  | 10-DEC-2008  | 5         |
 8 | BBBBB  | 11-DEC-2008  | 2         |
 9 | BBBBB  | 12-DEC-2008  | 0         |
10 | BBBBB  | 09-DEC-2009  | 3         |
11 | BBBBB  | 10-DEC-2009  | 1         |
12 | BBBBB  | 11-DEC-2009  | 0         |
   |========|==============|===========|

我需要把它转换成一个表格。

   |========|==============|===============|===============|
   | Name   | Date         | Quantity 2008 | Quantity 2009 |
   |========|==============|===============|===============|
 1 | AAAAA  | 09-DEC       |               | 3             |
 2 | AAAAA  | 10-DEC       | 5             | 2             |
 3 | AAAAA  | 11-DEC       | 2             | 3             |
 4 | AAAAA  | 12-DEC       | 0             |               |
 5 | BBBBB  | 09-DEC       |               | 3             |
 6 | BBBBB  | 10-DEC       | 5             | 1             |
 7 | BBBBB  | 11-DEC       | 2             | 0             |
 8 | BBBBB  | 12-DEC       | 0             |               |
   |========|==============|===============|===============|

应该假设数据库中有数千个命名项目，并且每年记录的开始日期会有所不同，例如，2010 年 12 月 6 日和 2009 年 12 月 5 日。

希望有人能提供帮助。

Answer 1

如果您使用的是 Oracle 11g，您可以将表格旋转为：

select *
  from (    
    select name,
           to_char(theDate, 'DD-MON') as dayMonth,
           extract(year from thedate) as year,
           quantity
      from yourTable
) 
 pivot (sum(quantity) as quantity for (year) in (2008 as y2008, 2009 as y2009))
order by name, dayMonth;

返回以下内容

NAME  DAYMONTH        Y2008_QUANTITY         Y2009_QUANTITY         
----- --------------- ---------------------- ---------------------- 
AAAAA 09-DEC                                 3                      
AAAAA 10-DEC          5                      2                      
AAAAA 11-DEC          2                      3                      
AAAAA 12-DEC          0                                        
BBBBB 09-DEC                                 3                      
BBBBB 10-DEC          5                      1                      
BBBBB 11-DEC          2                      0                      
BBBBB 12-DEC          0                                           

8 rows selected

Answer 2

在 Oracle11 之前，您可以使用这个标准的数据透视技巧：

SQL> create table mytable (name,mydate,quantity)
  2  as
  3  select 'AAAAA', date '2008-12-10', 5 from dual union all
  4  select 'AAAAA', date '2008-12-11', 2 from dual union all
  5  select 'AAAAA', date '2008-12-12', 0 from dual union all
  6  select 'AAAAA', date '2009-12-09', 3 from dual union all
  7  select 'AAAAA', date '2009-12-10', 2 from dual union all
  8  select 'AAAAA', date '2009-12-11', 3 from dual union all
  9  select 'BBBBB', date '2008-12-10', 5 from dual union all
 10  select 'BBBBB', date '2008-12-11', 2 from dual union all
 11  select 'BBBBB', date '2008-12-12', 0 from dual union all
 12  select 'BBBBB', date '2009-12-09', 3 from dual union all
 13  select 'BBBBB', date '2009-12-10', 1 from dual union all
 14  select 'BBBBB', date '2009-12-11', 0 from dual
 15  /

Table created.

SQL> select name
  2       , to_char(mydate,'dd-MON')
  3       , sum(case extract(year from mydate) when 2008 then quantity end) quantity_2008
  4       , sum(case extract(year from mydate) when 2009 then quantity end) quantity_2009
  5    from mytable
  6   group by name
  7       , to_char(mydate,'dd-MON')
  8   order by name
  9       , to_char(mydate,'dd-MON')
 10  /

NAME  TO_CHA QUANTITY_2008 QUANTITY_2009
----- ------ ------------- -------------
AAAAA 09-DEC                           3
AAAAA 10-DEC             5             2
AAAAA 11-DEC             2             3
AAAAA 12-DEC             0
BBBBB 09-DEC                           3
BBBBB 10-DEC             5             1
BBBBB 11-DEC             2             0
BBBBB 12-DEC             0

8 rows selected.

问候，
抢。

Answer 3

这样的事情应该可以完成：

SELECT
    Name, FormattedDate, SUM(Quantity_2008), SUM(Quantity_2009)
FROM
(
SELECT
    Name,
    ConvertDateToDayMonthRepresentation(YourDate) as FormattedDate,
    DECODE(EXTRACT(YEAR FROM YourDate), 2008, Quantity, 0) as Quantity_2008, 
    DECODE(EXTRACT(YEAR FROM YourDate), 2009, Quantity, 0) as Quantity_2009
FROM
    YourTable
)
GROUP BY Name, FormattedDate;

请注意：

1. 您需要为想要拥有的每一年添加代码 - 您的查询不能有动态列。
2. ConvertDateToDayMonthRepresentation 是一个用户定义的函数，它应该将日期转换为没有年份的日期和月份的字符串表示形式。使用EXTRACT 创建很简单。