将多行电话数据合并为显示开始和结束日期的单行的脚本

Question

对于下表，我想：

一世。 合并行，以便对于每种电话类型，将相同电话号码的集合合并为具有单个开始日期和结束日期的唯一电话号码

ii. 合并电话号码的 START DATE 是集合的最早开始日期，合并电话号码的 END DATE 是集合的 LATEST 结束日期

iii. 此外，我想确保当某些不同的电话号码按时间顺序夹在一组相同的电话号码中（由红色虚线箭头表示）时，那些相同的电话号码不会合并到一行中。 相反，它们被合并为单独的集合。 这是因为按时间顺序，它们现在是独立的集合

到目前为止我使用的脚本满足第一个和第二个要求，但不满足第三个要求

SELECT CLAIMANT_ID, PHONE_TYPE, PHONE_NUMBER, 
Min(START_DATE), Max(END_DATE) KEEP (DENSE_RANK FIRST ORDER BY END_DATE DESC NULLS FIRST)
FROM ClaimantData
GROUP BY CLAIMANT_ID, PHONE_TYPE, PHONE_NUMBER

有谁知道是否有任何方法可以增强脚本（或具有完全不同的脚本）也可以满足第三个要求？

预期结果如下：

插入的示例数据

CREATE TABLE CLAIMANTDATA (    
CLAIMANT_ID NUMBER(9) NOT NULL,
PHONE_TYPE VARCHAR2(50 BYTE) NOT NULL,
PHONE_NUMBER VARCHAR2(50 BYTE) NULL,
START_DATE DATE NOT NULL,
END_DATE DATE NULL
);

Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Primary',7653030303,to_date('12-JAN-12','DD-MON-RR'),to_date('23-JAN-12','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Primary',7653030303,to_date('09-JAN-12','DD-MON-RR'),to_date('12-JAN-12','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Primary',7653030303,to_date('01-JAN-12','DD-MON-RR'),to_date('09-JAN-12','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('09-MAR-11','DD-MON-RR'),to_date('01-JAN-12','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('22-FEB-11','DD-MON-RR'),to_date('09-MAR-11','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('19-AUG-10','DD-MON-RR'),to_date('22-FEB-11','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('16-AUG-10','DD-MON-RR'),to_date('19-AUG-10','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',2062009876,to_date('14-AUG-10','DD-MON-RR'),to_date('16-AUG-10','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('08-APR-10','DD-MON-RR'),to_date('14-AUG-10','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('06-APR-10','DD-MON-RR'),to_date('08-APR-10','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('08-JUL-09','DD-MON-RR'),to_date('06-APR-10','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('15-JAN-09','DD-MON-RR'),to_date('08-JUL-09','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7650880880,to_date('22-DEC-08','DD-MON-RR'),to_date('15-JAN-09','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('07-AUG-08','DD-MON-RR'),to_date('22-DEC-08','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('06-AUG-08','DD-MON-RR'),to_date('07-AUG-08','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('06-AUG-08','DD-MON-RR'),to_date('06-AUG-08','DD-MON-RR'));
Insert into CLAIMANTDATA (CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,START_DATE,END_DATE) values (504,'Alternate',7651000001,to_date('13-NOV-07','DD-MON-RR'),to_date('28-MAR-08','DD-MON-RR'));

Answer 1

这称为孤岛和间隙问题。

您可以使用解析 function 如下：

Select CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER,
        Min(start_date),
        Max(end_date)
  From
(Select t.*,
       Row_number() over (partition by CLAIMANT_ID, PHONE_TYPE order by start_date) as rn,
       Row_number() over (partition by CLAIMANT_ID, PHONE_TYPE, PHONE_NUMBER order by start_date) as rn_p
   From CLAIMANTDATA t) t
Group by CLAIMANT_ID,PHONE_TYPE,PHONE_NUMBER, rn-rn_p

Answer 2

在 Oracle 12.1 或更高版本中，解决此类间隙和孤岛问题的最简单方法是使用match_recognize子句。

select claimant_id, phone_type, phone_number, start_date, end_date
from   claimantdata
match_recognize (
  partition by claimant_id, phone_type
  order     by start_date
  measures  a.phone_number as phone_number,
            a.start_date as start_date, last(end_date) as end_date
  pattern   ( a b* )
  define    b as phone_number = a.phone_number and start_date = prev(end_date)
);