对任何可以提供帮助的人，
我想在mysql中创建一个数据查询，以递增编号按事务ID分组的行，并在下一个事务ID上重新开始编号，依此类推。
这是我的数据样本：
transaction_id | run_index
--------------------------
1001 | 20
1001 | 25
1001 | 31
1001 | 40
1002 | 4
1002 | 11
1002 | 16
1002 | 21
1002 | 35
1003 | 07
1003 | 18
1003 | 44
我希望它看起来像这样：
transaction_id | run_index | run_order
--------------------------------------
1001 | 20 | 1
1001 | 25 | 2
1001 | 31 | 3
1001 | 40 | 4
1002 | 4 | 1
1002 | 11 | 2
1002 | 16 | 3
1002 | 21 | 4
1002 | 35 | 5
1003 | 7 | 1
1003 | 18 | 2
1003 | 44 | 3
这是针对数据分析的特殊报告。
答案 0 :(得分：0)
SELECT transaction_id, run_index,
CASE WHEN @transaction_id != transaction_id THEN (@run := 1) - !(@transaction_id := transaction_id) ELSE @run := @run + 1 END AS run_order
FROM ( SELECT transaction_id, run_index
FROM table_name, (SELECT @transaction_id := NULL, @run := NULL) AS var
ORDER by transaction_id ASC, run_index DESC) AS h
应该是你想要的。
答案 1 :(得分：0)
SELECT transaction_id,
run_index,
IF(@CG=transaction_id, @I:=@I+1, @I:=1),
@CG:=transaction_id AS run_order
FROM ...
INNER JOIN (SELECT @I:=0, @CG:=-1) x
示例结果集：
+---------+--------------------------------+--------------+
| feed_id | IF(@CG=feed_id,@I:=@I+1,@I:=1) | @CG:=feed_id |
+---------+--------------------------------+--------------+
| 8 | 1 | 8 |
| 8 | 2 | 8 |
| 8 | 3 | 8 |
| 14 | 1 | 14 |
| 14 | 2 | 14 |
| 14 | 3 | 14 |
| 14 | 4 | 14 |
| 14 | 5 | 14 |
| 14 | 6 | 14 |
| 14 | 7 | 14 |
+---------+--------------------------------+--------------+