Question

图片我们有一个这样的表：

表1

+----------+----------+--------+------------+
| position | epoc     | name   | value      |
+----------+----------+--------+------------+
|    1     |    1     |   A    |        v01 |
|    1     |    1     |   B    |        v02 |
|    1     |    1     |   C    |        v03 |
|    1     |    2     |   A    |        v04 |
|    1     |    2     |   B    |        v05 |
|    1     |    2     |   C    |        v06 |
|    1     |    3     |   A    |        v07 |
|    1     |    3     |   B    |        v08 |
|    1     |    3     |   C    |        v09 |
|    1     |    4     |   A    |        v10 |
|    1     |    4     |   B    |        v11 |
|    1     |    4     |   C    |        v12 |
|    2     |    5     |   A    |        v13 |
|    2     |    5     |   B    |        v14 |
|    2     |    5     |   C    |        v15 |
|    2     |    6     |   A    |        v16 |
|    2     |    6     |   B    |        v17 |
|    2     |    6     |   C    |        v18 |
|    2     |    7     |   A    |        v19 |
|    2     |    7     |   B    |        v20 |
|    2     |    7     |   C    |        v21 |
|    2     |    8     |   A    |        v22 |
|    2     |    8     |   B    |        v23 |
|    2     |    8     |   C    |        v24 |
+----------+----------+--------+------------+

我希望能够得到这张表：

表2

+----------+--------------------+
| position | value              |
+----------+--------------------+
|    1     |    v01,v02,v04,v05 |
|    2     |    v13,v14,v16,v17 | 
+----------+--------------------+

条件是：

只是＆＃34;值＆＃34; ＆＃34; name＆＃34; A或B;
JUST＆＃34; epocs＆＃34;这是＆＃34;位置＆＃34;中的前2个独特结果（epoc 3,4,7,8被丢弃）
GROUP by table1 position（对于每个位置，我想要匹配先前条件的值的连接）

Answer 1

这可能就是你要找的东西：

select position, 
  group_concat(value order by overall_row_num) value
from
(
  select position,
      name,
      value,
      epoc,
       @num := if(@position = `position`, @num + 1, 1) as group_row_number,
       @position := `position` as dummy, 
      overall_row_num
  from
  (
    select position, name, 
      epoc, 
      value,
      @rn:=@rn+1 overall_row_num
    from t1, (SELECT @rn:=0) r
    where name in ('A', 'B')
    order by position, epoc
  ) x
  order by overall_row_num
) x1
where group_row_number <= 4
group by position

请参阅SQL Fiddle with demo

如何按/ group_concat分组每组结果？

1 个答案: