如何在Sql Server中使用GROUP BY作爲多列時選擇最長的字符串

Question

我有下表具有重複值的表。我只想從中獲取最大長度的文本。

+----+---------------------------------+-------------------+ 
| ID | Text       | no    | 
+----+---------------------------------+-------------------+ 
| 1 | lorem ipsum ipsum    | 8955\5445   | 
+----+---------------------------------+-------------------+ 
| 1 | ipsum       | 1879668\4554\4554 | 
+----+---------------------------------+-------------------+ 
| 1 | lorem ipsum ipsum    | 5464    | 
+----+---------------------------------+-------------------+ 
| 2 | lorem ipsum ipsum derome  | 13465465\54555\45 | 
+----+---------------------------------+-------------------+ 
| 2 | lorem ipsum derome ipsum derome | 555454   | 
+----+---------------------------------+-------------------+ 

我的查詢：

select id,MAX(text),MAX(no) from table1 group by id 

結果：

+----+--------------------------+-----------+ 
| id | Text      | no  | 
+----+--------------------------+-----------+ 
| 1 | lorem ipsum ipsum  | 8955\5445 | 
+----+--------------------------+-----------+ 
| 2 | lorem ipsum ipsum derome | 555454 | 
+----+--------------------------+-----------+ 

預期輸出：

+----+--------------------------+-------------------+ 
| id | Text      | no    | 
+----+--------------------------+-------------------+ 
| 1 | lorem ipsum ipsum  | 1879668\4554\4554 | 
+----+--------------------------+-------------------+ 
| 2 | lorem ipsum ipsum derome | 13465465\54555\45 | 
+----+--------------------------+-------------------+ 

Answer 1

你似乎想要最長的文本和最高的數字看你的預期結果。

讓我們來做一些測試數據;

IF OBJECT_ID('tempdb..#TestData') IS NOT NULL DROP TABLE #TestData 
GO 
CREATE TABLE #TestData (ID int, Text varchar(50), no int) 
INSERT INTO #TestData (ID, Text, no) 
VALUES 
(1,'lorem ipsum ipsum',8955) 
,(1,'ipsum',6879668) 
,(1,'lorem ipsum ipsum',5464) 
,(2,'lorem ipsum ipsum derome',63465465) 
,(2,'lorem ipsum derome ipsum derome',555454) 

我已經做了這個子查詢來分別獲得兩個最大值。第一個內部聯接僅用於篩選，第二個內部聯接返回no列中的最大值。

SELECT DISTINCT 
t.id 
,m.Text Text 
,MAX(t.no) Number 
FROM #TestData t 
LEFT JOIN --Get the max data length 
    (
     SELECT DISTINCT 
     t.ID 
     ,t.Text 
     FROM #TestData t 
     JOIN 
      (
       SELECT 
       ID 
       ,MAX(len(Text)) MaxLen 
       FROM #TestData 
       GROUP BY ID 
      ) mx 
     ON t.ID = mx.ID 
     AND LEN(t.Text) = mx.MaxLen 
    ) m 
ON t.ID = m.ID 
LEFT JOIN --Get the max no 
     (
     SELECT 
     ID 
     ,MAX(LEN(no)) MaxNo 
     FROM #TestData 
     GROUP BY ID 
    ) n 
ON t.ID = n.ID 
AND LEN(t.no) = n.MaxNo 
GROUP BY t.id, m.Text 

結果出來了;

id Text       MaxNo 
1 lorem ipsum ipsum    6879668 
2 lorem ipsum derome ipsum derome 63465465 

Answer 2

SELECT Id,Text,Number 
WHERE LENGTH(description) = (
      SELECT MAX(LENGTH(Text)) 
GROUP BY ID,Text,Number 

Answer 3

一個簡單的方法是做一個NOT EXISTS，即只要沒有其他具有相同ID的行具有更長的文本就返回一行。自我JOIN與GROUP BY，返回每個ID的最大號碼。

select t1.id, t1.Text, t3.no 
from tablename t1 
    join (select id, max(no) no from tablename group by id) t3 on t1.id = t3.id 
where not exists (select * from tablename t2 
        where t2.id = t1.id 
        and length(t2.Text) > length(t1.Text)) 

Answer 4

試試這個查詢，這個返回你需要什麼。 它的工作原理

SELECT a.id, 
     a.Text, 
     a.no 
FROM table AS a 
INNER JOIN 
(
    SELECT id, 
      MAX(LEN(Text)) AS lenght 
    FROM table 
    GROUP BY id 
) b 
    ON LEN(a.Text) = b.lenght 
     AND a.id = b.id; 

Answer 5

您可以將RANK函數與CTE結合使用。成纔是這樣的：

WITH cte AS (
SELECT 
    id 
    ,text 
    ,no 
    ,RANK() OVER(PARTITION BY id ORDER BY LEN(text) DESC) AS pos 
) 
SELECT * 
FROM cte 
WHERE pos = 1 

Answer 6

而且，可以通過下面的查詢，如果我們想只獲得最大的長度Text

Select top 1 id,Test,no --Top n, depending upon requirement 
From 
(Select id, Test, no, LEN(TEST) as Length 
From #TableName)B 
Order by Length Desc;