Question

我有一个表搜索，列是：

id(int), title(nvarchar), count(int), expression(nvarchar)

值：

1, aaa, 5, car 
2, aaa, 5, car poster
3, bbb, 8, car magazine
4, bbb, 8, car tv
5, ccc, 12, boat
6, ccc, 12 , boat tv

我需要以下结果

4, bbb, 8, car tv
3, bbb, 8, car magazine
2, aaa, 5, car poster
1, aaa, 5, car

或更好的结果来自4,3和2,1 像这样：

3, bbb, 8, car magazine
2, aaa, 5, car poster

我试过了：

SELECT   title 
FROM 
    (SELECT * 
     FROM Search 
     ORDER BY count DESC) 
WHERE expression 
LIKE '%car%'
GROUP BY  title

我还尝试过DISTINCT

Answer 1

使用带有CTE功能的ROW_NUMBER来查找匹配的行，并按降序计数和绑定断路器的ID对它们进行排序。然后选择每个分区（标题）中的第一个编号行。

WITH cteFindCar AS (
    SELECT id, title, count, expression,
           ROW_NUMBER() OVER(PARTITION BY title ORDER BY count DESC, id) AS RowNum
        FROM search
        WHERE expression LIKE '%car%'
)
SELECT id, title, count, expression
    FROM cteFindCar
    WHERE RowNum = 1;

SQL按顺序选择按顺序分组

1 个答案: