Question

我有一个查询可以提供正确的信息，但我无法弄清楚如何获得不同的日子并总结计数。下面是我的查询，结果显示我得到了什么，以及我想回来的内容。我以为工会会加入他们并给我我想要的结果。我已经搜索了很长时间，并没有找到任何关于这个主题的帮助。任何帮助将不胜感激。提前谢谢。

SELECT COUNT(log_datetime) AS icount, CONVERT(varchar, log_datetime, 101) AS logdate
from openrowset('sqloledb', 'ServerName1';'UserID';'Password',
          'select * from DatabaseName..TableName where field1 > 899')
group by convert(varchar, log_datetime, 101)
union
SELECT COUNT(log_datetime) AS icount, CONVERT(varchar, log_datetime, 101) AS logdate
from openrowset('sqloledb', 'ServerName2';'UserID';'Password',
    'select * from DatabaseName..TableName where field1 > 899')
group by convert(varchar, log_datetime, 101)
order by logdate

Results
235     01/10/2013
312     01/10/2013
3091    01/11/2013
3197    01/11/2013
3339    01/12/2013
3536    01/12/2013

Wanted Results
547      01/10/2013
6288     01/11/2013
6875     01/12/2013

Answer 1

你可以试试这个

SELECT SUM(icount) AS icount, logdate FROM
(
    SELECT COUNT(log_datetime) AS icount, CONVERT(varchar, log_datetime, 101) AS logdate
    FROM openrowset('sqloledb', 'ServerName1';'UserID';'Password',
              'select * from DatabaseName..TableName where field1 > 899')
    GROUP BY convert(varchar, log_datetime, 101)
        UNION ALL
    SELECT COUNT(log_datetime) AS icount, CONVERT(varchar, log_datetime, 101) AS logdate
    FROM openrowset('sqloledb', 'ServerName2';'UserID';'Password',
        'select * from DatabaseName..TableName where field1 > 899')
    GROUP BY convert(varchar, log_datetime, 101)
) AS tbl
GROUP BY logdate
ORDER BY logdate

这是sqlfiddle

Answer 2

SELECT count(1) icount, AA.logdate FROM
(
SELECT CONVERT(varchar, log_datetime, 101) AS logdate
from openrowset('sqloledb', 'ServerName1';'UserID';'Password',
          'select * from DatabaseName..TableName where field1 > 899')
union all
SELECT CONVERT(varchar, log_datetime, 101) AS logdate
from openrowset('sqloledb', 'ServerName2';'UserID';'Password',
    'select * from DatabaseName..TableName where field1 > 899')
) AA
group by AA.logdate order by AA.logdate

Answer 3

我相信peterm的答案是正确的，只是想添加一个关于原因的说明。

使用SQL，你必须在集合中思考。您正在创建两个唯一的数据集，并使用UNION将它们组合在一起。但是，GROUP BY语句在每个集合中，而不是组合。因此，多行结果。通过使用另一个SELECT语句包装所有内容，您将查询组合集，现在可以进行分组和聚合。

从2台机器和组结果中查询2个表

3 个答案: