在SQL Server中联接多个表

Question

我有一条select语句，该语句必须连接多个表，包括sys.indexes，sys.objects，sys.schemas和sys.dm_db_index_physical_stats，以获得这样的表

这是声明：

select
    db_name() as [Database],
    sc.[name] as [Schema],
    obj.[name] as [Table],  
    idx.[type_desc] as [Index Type],
    idx.[name] as [Index Name],
    ips.[avg_fragmentation_in_percent] as [Fragmentation Percent]
from 
    sys.indexes as idx
inner join 
    sys.objects as obj on idx.object_id = obj.object_id
inner join 
    sys.schemas as sc  on obj.schema_id = sc.schema_id
inner join 
    sys.dm_db_index_physical_stats( NULL,NULL, NULL, NULL ,'LIMITED') AS ips on obj.object_id = ips.object_id 
where 
    idx.[name] is not null 
    and obj.[type] = 'u' or obj.[type] = 'v' 
order by 
    [Fragmentation Percent] desc

问题是该语句工作不正确，并且两次包含一些行（这样做像是交叉联接，但不完全是这样）。谁能指出我的错误？

感谢您的时间！

Answer 1

使用CROSS_APPLY代替INNER JOIN。这将允许您将相关值传递给表值函数。在此处将分区号添加到结果中：

SELECT
    db_name() as [Database],
    sc.[name] as [Schema],
    obj.[name] as [Table],  
    idx.[type_desc] as [Index Type],
    idx.[name] as [Index Name],
    ips.partition_number AS [Partition Number],
    ips.[avg_fragmentation_in_percent] as [Fragmentation Percent]
FROM sys.indexes as idx
INNER JOIN sys.objects as obj on idx.object_id = obj.object_id
INNER JOIN sys.schemas as sc  on obj.schema_id = sc.schema_id
CROSS APPLY sys.dm_db_index_physical_stats( DB_ID(), idx.object_id, idx.index_id, NULL ,'LIMITED') AS ips
ORDER BY [Fragmentation Percent] desc;

原始查询的错误是您没有在sys.dm_db_index_physical_stats的联接中包括index_id，并且没有为当前数据库中的对象过滤（相同的object_id可能存在于不同的数据库中）。 WHERE子句是多余的，因此我将其删除。

出于碎片整理的目的，应排除小索引，因为这些小索引是从混合盘区中分配的，即使重新构建，也会报告高碎片。