Sql 在两列中选择具有相似数据的不同数据

我有一张桌子

使用此示例数据：

('Pune','Mumbai',128);
('Mumbai','Pune',128);
('Pune','Nashik',200);
('Nashik','Pune',200);
('Nashik','Mumbai',250);
('Nashik','Mumbai',250);

我只想选择一次城市组合，即从孟买浦那和孟买浦那只能选择一行

我试着使用self-join，但是没有用。所以，如果你能提供一个查询，请

---

提前感谢。

如果c2>c1，请使用not exists检查切换的c1/c2列是否存在切换组合：

select distinct c1, c2, dist
from XXXTEST t1 
where c1 < c2
  or  not exists (select * from XXXTEST t2
                  where t1.c1 = t2.c2 and t1.c2 = t2.c1);

执行为：

SQL>CREATE TABLE XXXTEST (C1 VARCHAR(10),C2 VARCHAR(10), dist integer);
SQL>INSERT INTO xxxtest VALUES ('Pune','Mumbai',128);
SQL>
SQL>INSERT INTO xxxtest VALUES ('Mumbai','Pune',128);
SQL>INSERT INTO xxxtest VALUES ('Pune','Nashik',200);
SQL>INSERT INTO xxxtest VALUES ('Nashik','Pune',200);
SQL>INSERT INTO xxxtest VALUES ('Nashik','Mumbai',250);
SQL>INSERT INTO xxxtest VALUES ('Nashik','Mumbai',250);
SQL>    select distinct c1, c2, dist
SQL&    from XXXTEST t1 
SQL&    where c1 < c2
SQL&      or  not exists (select * from XXXTEST t2
SQL&                      where t1.c1 = t2.c2 and t1.c2 = t2.c1);
C1         C2                dist
========== ========== ===========
Mumbai     Pune               128
Nashik     Mumbai             250
Nashik     Pune               200

                  3 rows found

分层方法：

另外，假设城市之间的距离是唯一的，我认为你可以简单地选择不同的距离，而不是将其与城市及其组合复杂化

分析方法：

---

您可以按字母顺序对城市进行排序，然后使用简单的：

---

这个问题是不正确的。它将只为精确的副本提供一个唯一的行。完全不同于OP的要求。是的，当然。早上在这里太早了。。。我要一杯咖啡，然后编辑。

SQL>CREATE TABLE XXXTEST (C1 VARCHAR(10),C2 VARCHAR(10), dist integer);
SQL>INSERT INTO xxxtest VALUES ('Pune','Mumbai',128);
SQL>
SQL>INSERT INTO xxxtest VALUES ('Mumbai','Pune',128);
SQL>INSERT INTO xxxtest VALUES ('Pune','Nashik',200);
SQL>INSERT INTO xxxtest VALUES ('Nashik','Pune',200);
SQL>INSERT INTO xxxtest VALUES ('Nashik','Mumbai',250);
SQL>INSERT INTO xxxtest VALUES ('Nashik','Mumbai',250);
SQL>    select distinct c1, c2, dist
SQL&    from XXXTEST t1 
SQL&    where c1 < c2
SQL&      or  not exists (select * from XXXTEST t2
SQL&                      where t1.c1 = t2.c2 and t1.c2 = t2.c1);
C1         C2                dist
========== ========== ===========
Mumbai     Pune               128
Nashik     Mumbai             250
Nashik     Pune               200

                  3 rows found

SQL> SELECT DISTINCT c1,
  2    c2,
  3    dist
  4  FROM
  5    ( SELECT c1, c2, dist FROM xxxtest CONNECT BY nocycle prior c1 = c2
  6    )
  7  WHERE c1 > c2
  8  /

C1         C2               DIST
---------- ---------- ----------
Pune       Mumbai            128
Nashik     Mumbai            250
Pune       Nashik            200

SQL>

SQL> WITH DATA AS
  2    (SELECT c1,
  3      c2,
  4      dist,
  5      row_number() over(partition BY dist order by dist) rn
  6    FROM xxxtest
  7    )
  8  SELECT c1, c2, dist FROM DATA WHERE rn = 1
  9  /

C1         C2               DIST
---------- ---------- ----------
Pune       Mumbai            128
Pune       Nashik            200
Nashik     Mumbai            250

SQL>

select distinct 
   least(c1, c2) as c1, 
   greatest(c1,c2) as c2,
   dist
from XXXTEST