我正在寻找将以下数据转换为所需输出的帮助。我们有项目,LOC DAY级别的数据,需要转换为项目,LOC日期范围,以减少表中的记录数量和其他要求。
Item LOC RP_DATE RP_IND
1003785256 543 2016-11-05 Y
1003785256 543 2016-11-06 Y
1003785256 543 2016-11-07 Y
1003785256 543 2016-11-09 Y
1003785256 543 2016-11-10 Y
1003790365 150 2016-11-05 Y
1003797790 224 2016-11-05 Y
1003797790 224 2016-11-06 Y
1003797790 224 2016-11-07 Y
1003797790 224 2016-11-08 Y
所需输出:
Item LOC RP_ST_DATE RP_END_DATE
1003785256 543 2016-11-05 2016-11-07
1003785256 543 2016-11-09 2016-11-10
1003790365 150 2016-11-05 2016-11-05
1003797790 224 2016-11-05 2016-11-08
这种方法适用于MySQL。它使用有序子查询中的组合变量为每个"范围"建立公共开始日期。CROSS JOIN 仅用于初始化变量,它不会改变行数。一旦建立了共同的开始日期,它就会在外部查询中成为按查询分组的简单组。
SELECT Item, LOC, RP_IND, dr_begin, MAX(RP_DATE) dr_end
FROM (
SELECT
mytable.*
, @fin := CONVERT(IF(@item<=>item AND @loc<=>loc AND DATEDIFF(rp_date, @d)=1, @fin, rp_date), DATE) AS dr_begin
, @item := item
, @loc := loc
, @d := rp_date
FROM mytable CROSS JOIN (SELECT @item:=NULL, @loc:=NULL, @d:=NULL, @fin := NULL) AS init
ORDER BY item, loc, rp_date
) d
GROUP BY Item, LOC, RP_IND, dr_begin
;
+----+------------+-----+--------+------------+---------------------+
| | Item | LOC | RP_IND | dr_begin | dr_end |
+----+------------+-----+--------+------------+---------------------+
| 1 | 1003785256 | 543 | Y | 2016-11-05 | 07.11.2016 00:00:00 |
| 2 | 1003785256 | 543 | Y | 2016-11-09 | 10.11.2016 00:00:00 |
| 3 | 1003790365 | 150 | Y | 2016-11-05 | 05.11.2016 00:00:00 |
| 4 | 1003797790 | 224 | Y | 2016-11-05 | 08.11.2016 00:00:00 |
+----+------------+-----+--------+------------+---------------------+
注意<=>如果两个操作数均为 NULL 则返回 1
请参阅以下位置的查询:http://rextester.com/SEYG96251
#drop table mytable;
CREATE TABLE mytable(
Item INTEGER NOT NULL
,LOC INTEGER NOT NULL
,RP_DATE DATE NOT NULL
,RP_IND VARCHAR(1) NOT NULL
);
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003785256,543,'2016-11-05','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003785256,543,'2016-11-06','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003785256,543,'2016-11-07','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003785256,543,'2016-11-09','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003785256,543,'2016-11-10','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003790365,150,'2016-11-05','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003797790,224,'2016-11-05','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003797790,224,'2016-11-06','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003797790,224,'2016-11-07','Y');
INSERT INTO mytable(Item,LOC,RP_DATE,RP_IND) VALUES (1003797790,224,'2016-11-08','Y');