检查SQL Server 2005的索引密度和碎片信息(转)

Posted 2020-07-28 TwinStudio

   查询数据库中所有表的索引密度和碎片信息，以便为索引的重建和整理提供依据，也可以参考DBCC SHOWCONTIG，通常FRAGMENTATIOIN在30%以上建议重建，否则建议整理

SELECT i.name                           AS indexname,
         o.name                           AS tablename,
         s.name                           AS schemaname,
         f.index_type_desc                AS indextype,
         f.avg_page_space_used_in_percent AS indexdensity,
         f.avg_fragmentation_in_percent   AS indexfragmentation,
         f.page_count                     AS pages
    FROM sys.dm_db_index_physical_stats(DB_ID(), NULL, NULL, NULL, ‘SAMPLED‘) f
         INNER JOIN sys.objects o ON o.object_id = f.object_id
         INNER JOIN sys.schemas s ON o.schema_id = s.schema_id
         INNER JOIN sys.indexes i ON f.object_id = i.object_id AND f.index_id = i.index_id
   WHERE page_count > 50
     AND f.index_id > 0
ORDER BY o.name, i.index_id

LIMITED | SAMPLED | DETAILED | NULL | DEFAULT
     这些模式影响了如何收集碎片数据。LIMITED模式扫描堆所有的页，但对于索引，则只扫描叶级上面的父级别页。SAMPLED收集在堆或索引中1%采样 率的数据。DETAILED模式扫描所有页（堆或索引）。DETAILED是执行最慢的，但也是最精确的选项。指定NULL或DEFAULT的效果与 LIMITED模式的相同。