选择包含特定列中前 N 个唯一值的行

Question

For example, how do I select the rows that contain the first N=3 unique values in column X1?

Observe the first 3 unique values in column X1 are apple, car, and egg. How do I only select the rows that contain one of these values in column X1?

+----+---------+--------------+---------+
| X1      | X2                | X3      |
+---------+-------------------+---------+
| apple   | rob@hotmail.com   | 285     |
| apple   | geo@gmail.com     | 862     |
| car     | p6346@live.com    | 381     |
| egg     | simon@hotmail.com | -1058   |
| egg     | pierr@hotmail.com | 652     |
| egg     | j5@gmail.com      | 27      |
| grape   | peter@outlook.com | -1502   |
| grape   | ann@aol.com       | 621     |
| lime    | frank@gmail.com   | 501     |
| lime    | george@aol.com    | 314     |
| lime    | sam@gmail.com     | 615     |
| melon   | mike@hotmail.com  | 271     |
| melon   | jo@hotmail.com    | -97     |
| pear    | james@aol.com     | -97     |
+---------+-------------------+---------+

The desired result would be:

+----+---------+--------------+---------+
| X1      | X2                | X3      |
+---------+-------------------+---------+
| apple   | rob@hotmail.com   | 285     |
| apple   | geo@gmail.com     | 862     |
| car     | p6346@live.com    | 381     |
| egg     | simon@hotmail.com | -1058   |
| egg     | pierr@hotmail.com | 652     |
| egg     | j5@gmail.com      | 27      |
+---------+-------------------+---------+

Note X1 is ordered alphabetically.

Answer 1

You can use DENSE_RANK to achieve your required output as below-

DEMO HERE

 SELECT * FROM 
 (
      SELECT *,
      DENSE_RANK() OVER(ORDER BY [X1]) RN
      FROM your_table
 )A
 WHERE RN <= 3

Answer 2

Use a derived table for which you get the top 3 distinct x1 and join it to the original table on matching x1 s.

SELECT t.*
       FROM (SELECT DISTINCT
                    TOP (3)
                    ti.x1
                    FROM elbat ti
                    ORDER BY ti.x1) x
            INNER JOIN elbat t
                       ON t.x1 = x.x1;

Alternatively you can use the dense_rank() window function to assign a number to each record according to the order of its x1 .

SELECT x.x1,
       x.x2,
       x.x3
       FROM (SELECT ti.x1,
                    ti.x2,
                    ti.x3,
                    dense_rank() OVER (ORDER BY ti.x1) rk
                    FROM elbat ti) x
       WHERE x.rk <= 3;