Question

我想从以下记录中基于ID_DATE列获取每个组中的最高值。按ID_TOPIC

分组

CREATE TABLE DA_TBL(
    DATA_ID  VARCHAR2(50),
    REF_DESC VARCHAR2(50),
    DATE_L DATE NOT NULL,
    ID_TOPIC  VARCHAR2(50),
    ID_DATE NUMBER
);

INSERT all 
INTO DA_TBL  VALUES  ('1','sample 1',CURRENT_TIMESTAMP, 'local', 1)
INTO DA_TBL  VALUES  ('2','sample 2',CURRENT_TIMESTAMP, 'tradition', 2)
INTO DA_TBL  VALUES  ('3','sample 2',CURRENT_TIMESTAMP, 'gospel', 3)
INTO DA_TBL  VALUES  ('4','sample 4',CURRENT_TIMESTAMP, 'local', 4)
INTO DA_TBL  VALUES  ('5','sample 5',CURRENT_TIMESTAMP, 'gospel', 5)
INTO DA_TBL  VALUES  ('6','sample 6',CURRENT_TIMESTAMP, 'tradition', 6)
INTO DA_TBL  VALUES  ('7','sample 7',CURRENT_TIMESTAMP, 'gospel', 7)
INTO DA_TBL  VALUES  ('8','sample 8',CURRENT_TIMESTAMP, 'local', 8)
INTO DA_TBL  VALUES  ('9','sample 9',CURRENT_TIMESTAMP, 'tradition', 9)
INTO DA_TBL  VALUES  ('10','sample 10',CURRENT_TIMESTAMP, 'local', 10)
INTO DA_TBL  VALUES  ('11','sample 11',CURRENT_TIMESTAMP, 'gospel', 11)
SELECT * FROM dual;

我想要的是：

DATA_ID|REF_DESC |ID_TOPIC |ROWNUMBER|
-------|---------|---------|---------|
9      |sample 9 |tradition|        1|
10     |sample 10|local    |        1|
11     |sample 11|gospel   |        1|

我得到的是：

DATA_ID|REF_DESC|ID_TOPIC |ROWNUMBER|
-------|--------|---------|---------|
9      |sample 9|tradition|        1|
8      |sample 8|local    |        1|
7      |sample 7|gospel   |        1|

我尝试过的

 SELECT *
  FROM (SELECT DATA_ID, REF_DESC, ID_TOPIC 
              , ROW_NUMBER() OVER (PARTITION BY ID_TOPIC ORDER BY DATA_ID DESC) AS rownumber
         FROM  DA_TBL ORDER BY DATA_ID DESC)
 WHERE rownumber = 1;

Answer 1

您的问题是DATA_ID列的数据类型。它是一个字符串，并且字符串的排序方式不同于数字。 9比10“大”。

如果可以，请将列的数据类型修改为NUMBER。

或者在此处应用TO_NUMBER：

ROW_NUMBER() OVER (PARTITION BY ID_TOPIC ORDER BY TO_NUMBER(DATA_ID) DESC)
                                                  ------------------
                                                  here

如果列包含字母数字值，则此方法可能无效。

SQL> l
  1  SELECT *
  2    FROM (SELECT DATA_ID, REF_DESC, ID_TOPIC
  3                , ROW_NUMBER() OVER (PARTITION BY ID_TOPIC ORDER BY to_number(DATA_ID) DESC) AS rownumber
  4           FROM  DA_TBL ORDER BY DATA_ID DESC)
  5*  WHERE rownumber = 1
SQL> /

DATA_ID REF_DESC   ID_TOPIC    ROWNUMBER
------- ---------- ---------- ----------
9       sample 9   tradition           1
11      sample 11  gospel              1
10      sample 10  local               1

SQL>

Answer 2

尝试以下查询：

SELECT 
  t1.DATA_ID,
  t1.REF_DESC,
  t1.ID_TOPIC
FROM yourTable t1
JOIN (
  SELECT 
    ID_TOPIC, 
    MAX(ID_DATE) AS maxIdDate
  FROM yourTable 
  GROUP BY ID_TOPIC
) t2 ON t1.ID_TOPIC = t2.ID_TOPIC AND t1.ID_DATE = t2.maxIdDate

我认为您在查询中使用了错误的列，因此如果以这种方式修复它，它也应该起作用：

SELECT 
  *
FROM (
  SELECT 
    DATA_ID, 
    REF_DESC, 
    ID_TOPIC,
    ROW_NUMBER() OVER (PARTITION BY ID_TOPIC ORDER BY ID_DATE DESC) AS rownumber
  FROM DA_TBL
) t
WHERE rownumber = 1;

Answer 3

这是因为您在varchar上拥有DATA_ID数据类型。

要正确排序（您希望整数排序，而不是文本顺序），则需要进行强制转换：

ROW_NUMBER() OVER (PARTITION BY ID_TOPIC ORDER BY CAST(DATA_ID AS NUMBER) DESC)

旁注：表中的数据类型是错误的-如果您有ID，则自然应将其存储为数字（除非您有特殊要求）。

根据您发布的内容，您将两个ID列存储为varchar，而应该是number。

Answer 4

我想根据ID_DATE获得每个组中的最高价值

如果是这样，为什么您要专注于DATA_ID？

您想要的查询是：

SELECT *
FROM (SELECT DATA_ID, REF_DESC, ID_TOPIC ,
             ROW_NUMBER() OVER (PARTITION BY ID_TOPIC ORDER BY ID_DATE DESC) AS rownumber
      FROM  DA_TBL
     )
WHERE rownumber = 1
ORDER BY TO_NUMBER(DATA_ID) DESC;

Here是db <>小提琴。

从每个组的Oracle获取最大值

4 个答案: