Question

我想获得一个包含独特用户的Highscore列表。遇到独特部分的问题。通过删除GROUP BY，脚本不包含错误

数据库：

GameResult
Id, GameId, TicketId,   Score,  Email,          Extra,  CreatedTime
1   1       1           100     test@test.com   info1   2012-12-01T12:12:12
2   1       2           200     test2@test.com  info2   2014-01-01T01:01:01
3   1       3           300     test3@test.com  info3   2013-05-01T05:05:05

Ticket
Id, UserId, TicketNumber,   ControlCode
1   1       abc1            123
2   1       abc2            234
3   2       abc3            345

User
Id, UniqueUserId
1   555
2   666

QUERY：

SELECT 
  g1.*, t1.UserId, t1.ControlCode, t1.TicketNumber, u1.UniqueUserId
FROM [dbo].[GameResult] as g1
INNER JOIN [dbo].[Ticket] as t1
      ON (g1.TicketId = t1.Id)
INNER JOIN [dbo].[User] as u1
    ON (t1.UserId = u1.Id)
GROUP BY u1.UniqueUserId 
ORDER BY g1.Score Desc, g1.CreatedTime Asc

ERROR：

Column 'dbo.GameResult.Id' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause.

通缉结果：

Rank    GameResult.Id,  GameId, TicketId,   Score,  Email,          Extra,  CreatedTime,            UserId, TicketNumber,   ControlCode,    UniqueUserId
1       3               1       3           300     test3@test.com  info3   2013-05-01T05:05:05     2       abc3            345             666
2       2               1       2           200     test2@test.com  info2   2014-01-01T01:01:01     1       abc2            234             555

Answer 1

好的，如果我理解正确，您希望获得每个用户的最高分，并按分数降序返回查询。

在这种情况下，我会使用带有这样的Windows函数的CTE：

(\d{2})

如果有必要，可以在没有CTE和ROW_NUMBER函数的情况下完成所需的操作，但这将是一个更复杂的查询。这是一个更易于阅读的更清洁的解决方案，IMO。

Answer 2

您只需要在查询中添加rank即可。您可能希望使用dense_rank或row_number，具体取决于您希望如何处理具有相同分数的行。

这是一个模拟查询：

SELECT 
  rank() over (partition by g1.Id order by g1.Score Desc) as Rank,
  g1.Id,
  g1.GameId,
  g1.TicketId,
  g1.Score,
  g1.Email,
  g1.Extra,
  g1.CreatedTime,
  t1.UserId, 
  t1.ControlCode, 
  t1.TicketNumber, 
  u1.UniqueUserId
FROM [dbo].[GameResult] as g1
INNER JOIN [dbo].[Ticket] as t1
    ON (g1.TicketId = t1.Id)
INNER JOIN [dbo].[User] as u1
    ON (t1.UserId = u1.Id)
GROUP BY u1.UniqueUserId 
ORDER BY g1.Score Desc, g1.CreatedTime Asc

提供的排名查询基于示例数据，但很可能您必须计算每个游戏的排名，如下所示：

rank() over (partition by g1.GameId, g1.Id order by g1.Score Desc) as Rank,

Answer 3

使用内部联接获取有关SQL join的更多信息例子

SELECT * FROM GameResult as a 
    INNER JOIN Ticket as b
    ON a.TicketId=b.TicketId;
    INNER JOIN User as c
    ON b.UserId = c.Id

具有排名和独特用户的高分

3 个答案: