场景是对于特定日期范围，我需要返回字段值已更改的所有行（FROM和TO date）。然后，更改的字段值将以不同的颜色显示在屏幕上。
我可以返回特定时间段内发生变更的所有员工记录。但是，如果员工的记录在所选时间段内多次更改，我需要返回单个员工记录，其中包含每个不同X_FLAG列的组合值。 1表示发生了变化，0表示没有变化。
表DDL是：
CREATE TABLE "EMPLOYEE_DATA"
( "EMPLOYEE_ID" NUMBER(20,0),
"EMPLOYEE_NAME" VARCHAR2(100 BYTE),
"EMPLOYEE_NAME_FLAG" NUMBER(1,0),
"EMPLOYEE_ROLE" VARCHAR2(100 BYTE),
"EMPLOYEE_ROLE_FLAG" NUMBER(1,0),
"EMPLOYEE_SALARY" VARCHAR2(100 BYTE),
"EMPLOYEE_SALARY_FLAG" NUMBER(1,0),
"DATE_VALID_FROM" DATE,
"DATE_VALID_TO" DATE,
"HAS_RECORD_CHANGED" NUMBER(1,0),
"CURRENT_ROW_IND" NUMBER(1,0)
);
模拟数据是：
Insert into EMPLOYEE_DATA (EMPLOYEE_ID,EMPLOYEE_NAME,EMPLOYEE_NAME_FLAG,EMPLOYEE_ROLE,EMPLOYEE_ROLE_FLAG,EMPLOYEE_SALARY,EMPLOYEE_SALARY_FLAG, DATE_VALID_FROM,DATE_VALID_TO,HAS_RECORD_CHANGED,CURRENT_ROW_IND) values (1,'John Smith',0,'Associate',0,'1',0,to_date('01-FEB-17','DD-MON-RR'),to_date('28-FEB-17','DD-MON-RR'),0,0);
Insert into EMPLOYEE_DATA (EMPLOYEE_ID,EMPLOYEE_NAME,EMPLOYEE_NAME_FLAG,EMPLOYEE_ROLE,EMPLOYEE_ROLE_FLAG,EMPLOYEE_SALARY,EMPLOYEE_SALARY_FLAG, DATE_VALID_FROM,DATE_VALID_TO,HAS_RECORD_CHANGED,CURRENT_ROW_IND) values (2,'Katy Brown',0,'Team Leader',0,'7',0, to_date('01-FEB-17','DD-MON-RR'),to_date('28-FEB-17','DD-MON-RR'),0,0);
Insert into EMPLOYEE_DATA (EMPLOYEE_ID,EMPLOYEE_NAME,EMPLOYEE_NAME_FLAG,EMPLOYEE_ROLE,EMPLOYEE_ROLE_FLAG,EMPLOYEE_SALARY,EMPLOYEE_SALARY_FLAG, DATE_VALID_FROM,DATE_VALID_TO,HAS_RECORD_CHANGED,CURRENT_ROW_IND) values (2,'Katy Brown',0,'Team Leader',0,'7',0, to_date('01-APR-17','DD-MON-RR'),to_date('31-DEC-99','DD-MON-RR'),1,1);
Insert into EMPLOYEE_DATA (EMPLOYEE_ID,EMPLOYEE_NAME,EMPLOYEE_NAME_FLAG,EMPLOYEE_ROLE,EMPLOYEE_ROLE_FLAG,EMPLOYEE_SALARY,EMPLOYEE_SALARY_FLAG, DATE_VALID_FROM,DATE_VALID_TO,HAS_RECORD_CHANGED,CURRENT_ROW_IND) values (3,'Ian Jones',1,'Delivery Manager',1,'3',1, to_date('01-MAR-17','DD-MON-RR'),to_date('31-DEC-99','DD-MON-RR'),1,1);
Insert into EMPLOYEE_DATA (EMPLOYEE_ID,EMPLOYEE_NAME,EMPLOYEE_NAME_FLAG,EMPLOYEE_ROLE,EMPLOYEE_ROLE_FLAG,EMPLOYEE_SALARY,EMPLOYEE_SALARY_FLAG, DATE_VALID_FROM,DATE_VALID_TO,HAS_RECORD_CHANGED,CURRENT_ROW_IND) values (1,'John Smith',0,'Analyst',1,'1',0, to_date('01-MAR-17','DD-MON-RR'),to_date('31-MAR-17','DD-MON-RR'),1,0);
Insert into EMPLOYEE_DATA (EMPLOYEE_ID,EMPLOYEE_NAME,EMPLOYEE_NAME_FLAG,EMPLOYEE_ROLE,EMPLOYEE_ROLE_FLAG,EMPLOYEE_SALARY,EMPLOYEE_SALARY_FLAG, DATE_VALID_FROM,DATE_VALID_TO,HAS_RECORD_CHANGED,CURRENT_ROW_IND) values (1,'John Smith',0,'Analyst',0,'2',1, to_date('01-APR-17','DD-MON-RR'),to_date('31-DEC-99','DD-MON-RR'),1,1);
我的查询是：
SELECT *
FROM EMPLOYEE_DATA
WHERE DATE_VALID_FROM <= TO_DATE('01/04/2017', 'dd/mm/yyyy')
AND DATE_VALID_TO >= TO_DATE('01/01/2017', 'dd/mm/yyyy')
AND HAS_RECORD_CHANGED = '1'
ORDER BY EMPLOYEE_ID ASC, DATE_VALID_FROM ASC;
在我的最终查询中，我将添加以下行以恢复当前记录AND CURRENT_ROW_IND = '1'。我把它留下来展示我需要如何将“John Smith”记录的数据合并为合并EMPLOYEE_ROLE_FLAG和EMPLOYEE_SALARY_FLAG以用于前一个和当前记录（对于John Smith）
编辑 - 添加原始和目标结果。如果可能，我需要聚合以获得每个唯一员工的X_FLAG列的最大值。
原创
EMPLOYEE_ID EMPLOYEE_NAME EMPLOYEE_NAME_FLAG EMPLOYEE_ROLE EMPLOYEE_ROLE_FLAG EMPLOYEE_SALARY EMPLOYEE_SALARY_FLAG DATE_VALID_FROM DATE_VALID_TO HAS_RECORD_CHANGED CURRENT_ROW_IND
1 John Smith 0 Associate 0 1 0 01-Feb-17 28-Feb-17 0 0
2 Katy Brown 0 Team Leader 0 7 0 01-Feb-17 28-Feb-17 0 0
2 Katy Brown 0 Team Leader 0 7 0 01-Apr-17 31-Dec-99 1 1
3 Ian Jones 1 Delivery Manager 1 3 1 01-Mar-17 31-Dec-99 1 1
1 John Smith 0 Analyst 1 0 0 01-Mar-17 31-Mar-17 1 0
1 John Smith 0 Analyst 0 1 1 01-Apr-17 31-Dec-99 1 1
目标
EMPLOYEE_ID EMPLOYEE_NAME EMPLOYEE_NAME_FLAG EMPLOYEE_ROLE EMPLOYEE_ROLE_FLAG EMPLOYEE_SALARY EMPLOYEE_SALARY_FLAG DATE_VALID_FROM DATE_VALID_TO HAS_RECORD_CHANGED CURRENT_ROW_IND
1 John Smith 0 Analyst 1 1 1 01-Apr-17 31-Dec-99 1 1
2 Katy Brown 0 Team Leader 0 7 0 01-Apr-17 31-Dec-99 1 1
3 Ian Jones 1 Delivery Manager 1 3 1 01-Mar-17 31-Dec-99 1 1
答案 0 :(得分：1)
如果可能的话，我需要聚合以获得每个唯一员工的X_FLAG列的最大值。
对于上述要求，GROUP BY会起作用，对吗？
请找到以下查询以实现目标，
SELECT EMPLOYEE_ID
, EMPLOYEE_NAME
, MIN(EMPLOYEE_NAME_FLAG) EMPLOYEE_NAME_FLAG
, EMPLOYEE_ROLE
, MAX(EMPLOYEE_ROLE_FLAG) EMPLOYEE_ROLE_FLAG
, MIN(EMPLOYEE_SALARY) EMPLOYEE_SALARY
, MAX(EMPLOYEE_SALARY_FLAG) EMPLOYEE_SALARY_FLAG
, MAX(DATE_VALID_FROM) DATE_VALID_FROM
, MAX(DATE_VALID_TO) DATE_VALID_TO
, HAS_RECORD_CHANGED
, MAX(CURRENT_ROW_IND) CURRENT_ROW_IND
FROM EMPLOYEE_DATA
WHERE HAS_RECORD_CHANGED = 1
AND DATE_VALID_FROM BETWEEN TO_DATE('01/01/2017', 'dd/mm/yyyy')
AND TO_DATE('01/04/2017', 'dd/mm/yyyy')
GROUP BY EMPLOYEE_ID
, EMPLOYEE_NAME
, EMPLOYEE_ROLE
, HAS_RECORD_CHANGED
ORDER BY EMPLOYEE_ID ASC
, DATE_VALID_FROM ASC;
答案 1 :(得分：1)
考虑派生表，其中单元级别查询与计算最大标志的聚合查询连接：
SELECT emp.*
FROM
(SELECT *
FROM EMPLOYEE_DATA
WHERE DATE_VALID_FROM BETWEEN TO_DATE('01/01/2017', 'dd/mm/yyyy')
AND TO_DATE('01/04/2017', 'dd/mm/yyyy')
AND HAS_RECORD_CHANGED = 1
) AS emp
INNER JOIN
(SELECT EMPLOYEE_ID, MAX(EMPLOYEE_NAME_FLAG) AS MAX_NAME_FLAG,
MAX(EMPLOYEE_ROLE_FLAG) AS MAX_ROLE_FLAG,
MAX(EMPLOYEE_SALARY_FLAG) AS MAX_SALARY_FLAG
FROM EMPLOYEE_DATA
WHERE DATE_VALID_FROM BETWEEN TO_DATE('01/01/2017', 'dd/mm/yyyy')
AND TO_DATE('01/04/2017', 'dd/mm/yyyy')
AND HAS_RECORD_CHANGED = 1
GROUP BY EMPLOYEE_ID
) AS agg
ON emp.EMPLOYEE_ID = agg.EMPLOYEE_ID
AND emp.EMPLOYEE_NAME_FLAG = agg.MAX_NAME_FLAG
AND emp.EMPLOYEE_ROLE_FLAG = agg.MAX_ROLE_FLAG
AND emp.EMPLOYEE_SALARY_FLAG = agg.MAX_SALARY_FLAG
ORDER BY emp.EMPLOYEE_ID ASC, emp.DATE_VALID_FROM ASC