Question

我有一个程序，在firebird中生成星期五和星期四的日期。我使用以下查询生成数据列表：

    select w_start as "Friday", w_end as "Thursday",
(select count(*) from course C 
                         inner join enrolment E on E.cod_course=C.id
                         where E.date_enrolled between w_start and w_end
                                   and C_TYPE = 'PU' and CONFIRMED='C'  
                                  and (C.name like :PublicCourseOption1  or C.name like :PublicCourseOption2 or C.name like :PublicCourseOption3  or C.name like :PublicCourseOption4 
                                         or C.VERSION like :CourseVersion1 or C.version like :CourseVersion2  or C.version like :CourseVersion3 )
) as "Enrolments",
(select list(distinct promotype, ', ') from programmers where datesent between w_start and w_end) as "Promos", 
(select list(distinct course, ', ') from programmers where datesent between w_start and w_end) as "Courses"
from get_weeks(:dtFromDate, :dtToDate) wks
order by w_start

它会产生以下结果：

Friday      Thursday    Enrolments  Promos                Courses
04/01/2013  10/01/2013  5           FAX                   WHS
11/01/2013  17/01/2013  11          EMAIL                 WHS
18/01/2013  24/01/2013  6           FAX                   WHS
25/01/2013  31/01/2013  12          EMAIL, FAX            RTW, YSM103
01/02/2013  07/02/2013  17          EMAIL, FAX, Wcover    REF-CIT, WHS Toll, WorkCover
08/02/2013  14/02/2013  19          FAX                   HSR HUR- INFO
15/02/2013  21/02/2013  12          FAX                   MC
22/02/2013  28/02/2013  19          EMAIL, FAX            ARTW, DYS25, MC
01/03/2013  07/03/2013  22          COMCARE, FAX, Wcover  COMCARE, COMM, WorkCover
08/03/2013  14/03/2013  13          FAX                   HSR
15/03/2013  21/03/2013  12
22/03/2013  28/03/2013  16          FAX                   HSR

有谁知道如何将数据转换为数据透视表，如下所示：

Promos- Course - 10/01/2013 (Thursday Date)    - 17/01/2013 (Thursday Date)
FAX   -  WHS   - 15 enrolments                 - 25 enrolments
EMAIL -  MC    - 14 Enrolments                 - 36 enrolments

日期是动态创建的，并且每次运行查询时都会有所不同，因此静态查询无法满足我的需求。

Answer 1

德里克，

在我的工作中，我们遇到了类似的情况。我们的代码是从客户端执行的，所以我们最终做的是创建一个返回SQL语句的存储过程，然后我们执行返回的sql语句。

为了完成你的任务，我创建了一个名为X_GET_ENROLLMENTS_FOR_DATE的助手存储过程

CREATE OR ALTER PROCEDURE X_GET_ENROLLMENTS_FOR_DATE (
    PROMOS VARCHAR(100),
    COURSES VARCHAR(100),
    DATE_STRING VARCHAR(30))
RETURNS (
    ENROLLMENTS INTEGER)
AS
BEGIN

  SELECT SUM(ENROLLMENTS)
  FROM TABLE_X
  WHERE TABLE_X.PROMOS = :PROMOS
        AND TABLE_X.COURSES = :COURSES
        AND TABLE_X.THURSDAY = :DATE_STRING
  INTO ENROLLMENTS;

  SUSPEND;
END

我使用EXECUTE BLOCK创建了以下内容，但可以很容易地放在另一个存储过程中。

EXECUTE BLOCK
RETURNS (
    SQL VARCHAR(3000))
AS
DECLARE VARIABLE SELECT_SQL VARCHAR(2000);
DECLARE VARIABLE WHERE_SQL VARCHAR(2000);
DECLARE VARIABLE PROMOS VARCHAR(30);
DECLARE VARIABLE COURSES VARCHAR(30);
DECLARE VARIABLE THURSDAY VARCHAR(30);
DECLARE VARIABLE ENROLLMENTS INTEGER;
DECLARE VARIABLE FIELD_ID INTEGER; --DECLARE VARIABLE S varchar(1000)
DECLARE VARIABLE FIELD_COUNT INTEGER;
BEGIN
  SELECT_SQL = 'SELECT distinct promos, courses ,' || ASCII_CHAR(13);
  WHERE_SQL = 'FROM table_x' || ASCII_CHAR(13);

  FIELD_ID = 0;

  SELECT COUNT(DISTINCT THURSDAY)
  FROM TABLE_X
  INTO :FIELD_COUNT;

  FOR
  SELECT DISTINCT THURSDAY
  FROM TABLE_X
  INTO :THURSDAY
  DO
  BEGIN
    FIELD_ID = :FIELD_ID + 1;
    IF (:FIELD_ID = :FIELD_COUNT) THEN
      SELECT_SQL = :SELECT_SQL || 'T' || :FIELD_ID || '.Enrollments ' || 'Thu_' || REPLACE(:THURSDAY, '/', '_') || ASCII_CHAR(13);
    ELSE
      SELECT_SQL = :SELECT_SQL || 'T' || :FIELD_ID || '.Enrollments ' || 'Thu_' || REPLACE(:THURSDAY, '/', '_') || ',' || ASCII_CHAR(13);

    WHERE_SQL = :WHERE_SQL || 'LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, ''' || :THURSDAY || ''') T' || :FIELD_ID || ' on (1 = 1)' || ASCII_CHAR(13);
    --    WHERE_SQL = :WHERE_SQL || 'INNER JOIN ( Select Sum(Enrollments) Enrollments from table_x where promos = ''' || :PROMOS || ''' and Courses = ''' || :COURSES || ''' and thursday = ''' || :THURSDAY || ''') T' || :FIELD_ID || ' on (1 = 1)' || ASCII_CHAR(13);

  END
  SQL = :SELECT_SQL || :WHERE_SQL;
  SUSPEND;
END;

执行存储过程时，它将返回以下SQL，然后您可以执行该步骤。

SELECT distinct promos, courses ,
T1.Enrollments Thu_07_02_2013,
T2.Enrollments Thu_07_03_2013,
T3.Enrollments Thu_10_01_2013,
T4.Enrollments Thu_14_02_2013,
T5.Enrollments Thu_14_03_2013,
T6.Enrollments Thu_17_01_2013,
T7.Enrollments Thu_21_02_2013,
T8.Enrollments Thu_21_03_2013,
T9.Enrollments Thu_24_01_2013,
T10.Enrollments Thu_28_02_2013,
T11.Enrollments Thu_28_03_2013,
T12.Enrollments Thu_31_01_2013
FROM table_x
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '07/02/2013') T1 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '07/03/2013') T2 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '10/01/2013') T3 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '14/02/2013') T4 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '14/03/2013') T5 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '17/01/2013') T6 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '21/02/2013') T7 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '21/03/2013') T8 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '24/01/2013') T9 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '28/02/2013') T10 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '28/03/2013') T11 on (1 = 1)
LEFT OUTER JOIN X_GET_ENROLLMENTS_FOR_DATE ( table_x.PROMOS,table_x.COURSES, '31/01/2013') T12 on (1 = 1)

我想指出的一些事情...... 是我无法让列标题完全按照您的意愿来。此示例仅适用于星期四字段，您需要更改代码以处理星期五字段。我假设你可能有相同的日期，促销和课程的倍数，所以我在助手存储过程中总结了注册字段。

我希望这有帮助，

需要SQL Firebird Pivot表转换

1 个答案: