我试图弄清楚如何查询一个表（该表实际上是一个结果集，因此它将是一个子查询），将其分组ColA=ColB（见下文），并一步创建一个计算字段。
所以，如果我的测试数据看起来像
ColA ColB ColC
1 1个 1 2 bbbb 1 3 cccc 2 2 ddd 3 3 EEEE 3 4 ffff 3 5 格格 3 6 4 4 iii 5 5 6 6 咳咳 6 7 6 8 毫米
我只想检索其中的行，ColA=ColB并添加一个新列，告诉我是否ColA重复了原始数据。见下文。
ColA ColB ColC 倍数
1 1 aaaa 是 2 2 dddd 没有 3 3 eeee 是的 4 4 iii 没有 5 5 jjjj 没有 6 6 kkkk 是的
有人可以帮我解决语法吗？我一直在玩 Group By 和 SubSelects 无济于事。我是否需要对倍数字段使用 case 语句？
发布 create table 和 insert 语句而不是 Desc table 和 select * from table_name 更有帮助；http://tkyte.blogspot.com/2005/06/how-to-ask-questions.html
create table test_repeat(
cola number,
colb number,
colc varchar2(20)
);
insert into test_repeat values (1,1,'aaa');
insert into test_repeat values (1,2,'bbbb');
insert into test_repeat values (1,3,'cccc');
insert into test_repeat values (2,2,'dddd');
insert into test_repeat values (3,3,'eeee');
insert into test_repeat values (3,4,'ffff');
insert into test_repeat values (3,5,'gggg');
insert into test_repeat values (3,6,'hhhh');
insert into test_repeat values (4,4,'iiii');
insert into test_repeat values (5,5,'jjjj');
insert into test_repeat values (6,6,'kkkk');
insert into test_repeat values (6,7,'llll');
insert into test_repeat values (6,8,'mmmm');
commit;
1.您可以使用Oracle分析功能Lead 来查看您的结果集，看看colA是否与下一行（排序后..）相同，例如..
select * from
(select colA, colb,
(case when colA = (lead(cola) over
(partition by colA order by cola, colb))
then 'Yes'
else 'No'
end) multiples,
colc
from test_repeat)
where colA = colb
/
COLA COLB MUL COLC
---------- ---------- --- --------------------
1 1 Yes aaa
2 2 No dddd
3 3 Yes eeee
4 4 No iiii
5 5 No jjjj
6 6 Yes kkkk
2. 或者你可以获取COLA每个值的计数，比较一下是否有重复...
select a.colA, a.colb, a.colc,
(case when (select count(*) from test_repeat t where t.cola = a.colA) > 1
then 'Yes'
else 'No'
end) Repeat
from test_repeat a
where colA = colB
/
COLA COLB COLC REP
---------- ---------- -------------------- ---
1 1 aaa Yes
2 2 dddd No
3 3 eeee Yes
4 4 iiii No
5 5 jjjj No
6 6 kkkk Yes
它们都同样简单，但我建议使用分析函数方法，因为我发现它对于我过去使用过的所有查询通常更快。
SQL> select *
2 from test_repeat
3 order by cola
4 /
COLA COLB COLC
---------- ---------- --------------------
1 2 bbbb
1 1 aaa
1 3 cccc
2 2 dddd
3 4 ffff
3 3 eeee
3 5 gggg
3 6 hhhh
4 4 iiii
5 5 jjjj
6 6 kkkk
6 7 llll
6 8 mmmm
7 9 nnnn
14 rows selected.
SQL> select cola "ColA"
2 , max(decode(colb,cola,colb)) "ColB"
3 , max(decode(colb,cola,colc)) "ColC"
4 , case count(*) when 1 then 'no' else 'yes' end "multiples"
5 from test_repeat
6 group by cola
7 having cola = max(decode(colb,cola,colb))
8 order by cola
9 /
ColA ColB ColC mul
---------- ---------- -------------------- ---
1 1 aaa yes
2 2 dddd no
3 3 eeee yes
4 4 iiii no
5 5 jjjj no
6 6 kkkk yes
6 rows selected.
在性能方面，这个 SQL 大约等于 Rajesh 的第一个查询。所以你可以选择一个你更舒服的。
SQL> set autotrace on
SQL> select * from
2 (select colA, colb,
3 (case when colA = (lead(cola) over
4 (partition by colA order by cola, colb))
5 then 'Yes'
6 else 'No'
7 end) multiples,
8 colc
9 from test_repeat)
10 where colA = colb
11 /
COLA COLB MUL COLC
---------- ---------- --- --------------------
1 1 Yes aaa
2 2 No dddd
3 3 Yes eeee
4 4 No iiii
5 5 No jjjj
6 6 Yes kkkk
6 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 1491815685
-----------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 14 | 574 | 4 (25)| 00:00:01 |
|* 1 | VIEW | | 14 | 574 | 4 (25)| 00:00:01 |
| 2 | WINDOW SORT | | 14 | 532 | 4 (25)| 00:00:01 |
| 3 | TABLE ACCESS FULL| TEST_REPEAT | 14 | 532 | 3 (0)| 00:00:01 |
-----------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("COLA"="COLB")
Note
-----
- dynamic sampling used for this statement
Statistics
----------------------------------------------------------
4 recursive calls
0 db block gets
15 consistent gets
0 physical reads
0 redo size
421 bytes sent via SQL*Net to client
238 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
6 rows processed
SQL> /
COLA COLB MUL COLC
---------- ---------- --- --------------------
1 1 Yes aaa
2 2 No dddd
3 3 Yes eeee
4 4 No iiii
5 5 No jjjj
6 6 Yes kkkk
6 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 1491815685
-----------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 14 | 574 | 4 (25)| 00:00:01 |
|* 1 | VIEW | | 14 | 574 | 4 (25)| 00:00:01 |
| 2 | WINDOW SORT | | 14 | 532 | 4 (25)| 00:00:01 |
| 3 | TABLE ACCESS FULL| TEST_REPEAT | 14 | 532 | 3 (0)| 00:00:01 |
-----------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("COLA"="COLB")
Note
-----
- dynamic sampling used for this statement
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
7 consistent gets
0 physical reads
0 redo size
421 bytes sent via SQL*Net to client
238 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
6 rows processed
SQL> select cola "ColA"
2 , max(decode(colb,cola,colb)) "ColB"
3 , max(decode(colb,cola,colc)) "ColC"
4 , case count(*) when 1 then 'no' else 'yes' end "multiples"
5 from test_repeat
6 group by cola
7 having cola = max(decode(colb,cola,colb))
8 order by cola
9 /
ColA ColB ColC mul
---------- ---------- -------------------- ---
1 1 aaa yes
2 2 dddd no
3 3 eeee yes
4 4 iiii no
5 5 jjjj no
6 6 kkkk yes
6 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 3021378319
-----------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 14 | 532 | 4 (25)| 00:00:01 |
|* 1 | FILTER | | | | | |
| 2 | SORT GROUP BY | | 14 | 532 | 4 (25)| 00:00:01 |
| 3 | TABLE ACCESS FULL| TEST_REPEAT | 14 | 532 | 3 (0)| 00:00:01 |
-----------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("COLA"=MAX(DECODE("COLB","COLA","COLB")))
Note
-----
- dynamic sampling used for this statement
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
7 consistent gets
0 physical reads
0 redo size
421 bytes sent via SQL*Net to client
238 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
6 rows processed
SQL> /
ColA ColB ColC mul
---------- ---------- -------------------- ---
1 1 aaa yes
2 2 dddd no
3 3 eeee yes
4 4 iiii no
5 5 jjjj no
6 6 kkkk yes
6 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 3021378319
-----------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 14 | 532 | 4 (25)| 00:00:01 |
|* 1 | FILTER | | | | | |
| 2 | SORT GROUP BY | | 14 | 532 | 4 (25)| 00:00:01 |
| 3 | TABLE ACCESS FULL| TEST_REPEAT | 14 | 532 | 3 (0)| 00:00:01 |
-----------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("COLA"=MAX(DECODE("COLB","COLA","COLB")))
Note
-----
- dynamic sampling used for this statement
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
7 consistent gets
0 physical reads
0 redo size
420 bytes sent via SQL*Net to client
238 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
6 rows processed
SQL> set autotrace off
问候，
罗布。
我认为您需要该multiples列的子选择。不需要分组。写一些类似的东西：
select ColA, ColB, ColC,
(CASE
WHEN (select b.ColA from thistable b where b.ColA = a.ColA) > 1 THEN 'yes'
ELSE 'no'
END)
from thistable a
where ColA = ColB
我不知道领导功能，非常好。Rob 和 Rajesh，感谢到目前为止的回答，我想我能够使用两个子查询和一个外连接部分回答我自己的问题。
我猜我的查询不如其他人发布的那么有效，但我想知道人们是否可以对此查询发表评论。它会一直有效吗？是否可以对其进行调整以提高效率？它仍然需要为倍数输入“是”或“否”，现在它会将倍数计数或空值表示为 0。
SELECT BB.ColA, ColB, ColC, AA.CNT FROM (SELECT ColA, count(*) as CNT FROM TMPY where ColA<>ColB Group by ColA) AA, (SELECT ColA, ColB, ColC FROM TMPY WHERE COLA=COLB ) BB WHERE AA.COLa(+)=BB.COLB order by bb.ColA;
再次感谢你！