SQL Server数据库中所有索引和索引列的列表

Question

如何获取 SQL Server 2005+ 中所有索引和索引列的列表？我能找到的最接近的方法是：

select s.name, t.name, i.name, c.name from sys.tables t
inner join sys.schemas s on t.schema_id = s.schema_id
inner join sys.indexes i on i.object_id = t.object_id
inner join sys.index_columns ic on ic.object_id = t.object_id
inner join sys.columns c on c.object_id = t.object_id and
        ic.column_id = c.column_id

where i.index_id > 0    
 and i.type in (1, 2) -- clustered & nonclustered only
 and i.is_primary_key = 0 -- do not include PK indexes
 and i.is_unique_constraint = 0 -- do not include UQ
 and i.is_disabled = 0
 and i.is_hypothetical = 0
 and ic.key_ordinal > 0

order by ic.key_ordinal

这并不完全符合我的要求。
我想要的是列出所有用户定义的索引，（也就是不包括支持唯一约束和主键的索引），以及所有列（按照索引定义中的顺序排序），并尽可能提供更多元数据。

- Anton Gogolev

---

以上解决方案很优雅，但根据微软的说法，INDEXKEY_PROPERTY正在被弃用。请参见：http://msdn.microsoft.com/en-us/library/ms186773.aspx - Lisa

请注意，如果您要使用此处回答中的任何有效查询来编写索引脚本，则需要在查询中结合 sys.indexes 表的 filter_definition 列，以获取 SQL 2008+ AM 中非聚集索引的筛选器定义。 - user3101273

请注意，正如user3101273在下面指出的那样，所有答案都不包括索引过滤器（来自sys.indexes表的filter_definition列）。 - influent

Answer 1

您可以查看两个“sys”目录视图： sys.indexes 和 sys.index_columns。

这些视图可以提供关于索引及其列的几乎所有信息。

编辑：此查询已非常接近您要查找的内容：

SELECT 
     TableName = t.name,
     IndexName = ind.name,
     IndexId = ind.index_id,
     ColumnId = ic.index_column_id,
     ColumnName = col.name,
     ind.*,
     ic.*,
     col.* 
FROM 
     sys.indexes ind 
INNER JOIN 
     sys.index_columns ic ON  ind.object_id = ic.object_id and ind.index_id = ic.index_id 
INNER JOIN 
     sys.columns col ON ic.object_id = col.object_id and ic.column_id = col.column_id 
INNER JOIN 
     sys.tables t ON ind.object_id = t.object_id 
WHERE 
     ind.is_primary_key = 0 
     AND ind.is_unique = 0 
     AND ind.is_unique_constraint = 0 
     AND t.is_ms_shipped = 0 
ORDER BY 
     t.name, ind.name, ind.index_id, ic.is_included_column, ic.key_ordinal;

- marc_s

---

是的，我知道这些，但我无法安排所有必需的“sys.”目录，以便它们能够产生有意义的输出。 - Anton Gogolev

新版本好多了，但是"and ind.is_unique = 0"没有必要：它会过滤掉几乎所有需要的数据。然而，这个查询仍然包含太多的系统数据，我不知道如何去除它们。 - Anton Gogolev

@My-Name-Is: 该帖子作者想获取所有用户自定义的索引（is_ms_shipped=0），但不包括主键（is_primary_key=0）和仅用于支持唯一约束的索引（is_unique_constraint=0）。 - marc_s

@ZainRizvi 的 WHERE (1=1) 让他可以链式和重新排序他的 AND ... 而不用担心哪个是第一个。参见：https://dev59.com/CGsz5IYBdhLWcg3wCDl2#8149183 和 https://dev59.com/NnVC5IYBdhLWcg3wnCWA#242831 - basher

据我所知，index_column_id 不是主键列的顺序，而是表中的顺序！请使用 key_ordinal。 - Michel de Ruiter

Answer 2

您可以使用 sp_helpindex 命令来查看一个表的所有索引。

EXEC sys.sp_helpindex @objname = N'User' -- nvarchar(77)

对于所有的索引，你可以遍历 sys.objects 来获取每个表的所有索引。

- Dragon

---

问题在于它仅包含索引键列，而不包括所需列。 - John Odom

Answer 3

以上的方法都无法解决我的问题，但是这个可以：

-- KDF9's concise index list for SQL Server 2005+  (see below for 2000)
--   includes schemas and primary keys, in easy to read format
--   with unique, clustered, and all ascending/descendings in a single column
-- Needs simple manual add or delete to change maximum number of key columns
--   but is easy to understand and modify, with no UDFs or complex logic
--
SELECT
  schema_name(schema_id) as SchemaName, OBJECT_NAME(si.object_id) as TableName, si.name as IndexName,
  (CASE is_primary_key WHEN 1 THEN 'PK' ELSE '' END) as PK,
  (CASE is_unique WHEN 1 THEN '1' ELSE '0' END)+' '+
  (CASE si.type WHEN 1 THEN 'C' WHEN 3 THEN 'X' ELSE 'B' END)+' '+  -- B=basic, C=Clustered, X=XML
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,1,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,2,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,3,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,4,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,5,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  (CASE INDEXKEY_PROPERTY(si.object_id,index_id,6,'IsDescending') WHEN 0 THEN 'A' WHEN 1 THEN 'D' ELSE '' END)+
  '' as 'Type',
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,1) as Key1,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,2) as Key2,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,3) as Key3,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,4) as Key4,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,5) as Key5,
  INDEX_COL(schema_name(schema_id)+'.'+OBJECT_NAME(si.object_id),index_id,6) as Key6
FROM sys.indexes as si
LEFT JOIN sys.objects as so on so.object_id=si.object_id
WHERE index_id>0 -- omit the default heap
  and OBJECTPROPERTY(si.object_id,'IsMsShipped')=0 -- omit system tables
  and not (schema_name(schema_id)='dbo' and OBJECT_NAME(si.object_id)='sysdiagrams') -- omit sysdiagrams
ORDER BY SchemaName,TableName,IndexName

-------------------------------------------------------------------
-- or to generate creation scripts put a simple wrapper around that
SELECT SchemaName, TableName, IndexName,
  (CASE pk
    WHEN 'PK' THEN 'ALTER '+
     'TABLE '+SchemaName+'.'+TableName+' ADD CONSTRAINT '+IndexName+' PRIMARY KEY'+
     (CASE substring(Type,3,1) WHEN 'C' THEN ' CLUSTERED' ELSE '' END)
    ELSE 'CREATE '+
     (CASE substring(Type,1,1) WHEN '1' THEN 'UNIQUE ' ELSE '' END)+
     (CASE substring(Type,3,1) WHEN 'C' THEN 'CLUSTERED ' ELSE '' END)+
     'INDEX '+IndexName+' ON '+SchemaName+'.'+TableName
    END)+
  ' ('+
    (CASE WHEN Key1 is null THEN '' ELSE      Key1+(CASE substring(Type,4+1,1) WHEN 'D' THEN ' DESC' ELSE '' END) END)+
    (CASE WHEN Key2 is null THEN '' ELSE ', '+Key2+(CASE substring(Type,4+2,1) WHEN 'D' THEN ' DESC' ELSE '' END) END)+
    (CASE WHEN Key3 is null THEN '' ELSE ', '+Key3+(CASE substring(Type,4+3,1) WHEN 'D' THEN ' DESC' ELSE '' END) END)+
    (CASE WHEN Key4 is null THEN '' ELSE ', '+Key4+(CASE substring(Type,4+4,1) WHEN 'D' THEN ' DESC' ELSE '' END) END)+
    (CASE WHEN Key5 is null THEN '' ELSE ', '+Key5+(CASE substring(Type,4+5,1) WHEN 'D' THEN ' DESC' ELSE '' END) END)+
    (CASE WHEN Key6 is null THEN '' ELSE ', '+Key6+(CASE substring(Type,4+6,1) WHEN 'D' THEN ' DESC' ELSE '' END) END)+
    ')' as CreateIndex
FROM (
  ...
  ...listing SQL same as above minus the ORDER BY...
  ...
  ) as indexes
ORDER BY SchemaName,TableName,IndexName

----------------------------------------------------------
-- For SQL Server 2000 the following should work
--   change table names to sysindexes and sysobjects (no dots)
--   change object_id => id, index_id => indid,
--   change is_primary_key => (select count(constid) from sysconstraints as sc where sc.id=si.id and sc.status&15=1)
--   change is_unique => INDEXPROPERTY(si.id,si.name,'IsUnique')
--   change si.type => INDEXPROPERTY(si.id,si.name,'IsClustered')
--   remove all references to schemas including schema name qualifiers, and the XML type
--   add select where indid<255 and si.status&64=0 (to omit the text/image index and autostats)

如果你的名称中包含空格，在创建脚本中加上方括号。

当最后一个键列全部为空时，你就知道没有任何遗漏。

像原始请求一样过滤主键等是微不足道的。

注意：对于这个解决方案要小心，因为它不能区分索引列和包含列。

- KDF9

Answer 4

以下内容适用于SQL Server 2014/2016以及任何Microsoft Azure SQL Database。

生成一个全面的结果集，可轻松导出到Notepad/Excel进行切片和剖析，包括：

1. 表名
2. 索引名
3. 索引描述
4. 按顺序排列的索引列
5. 按顺序排列的包含列

 SELECT '[' + s.NAME + '].[' + o.NAME + ']' AS 'table_name'
    ,+ i.NAME AS 'index_name'
    ,LOWER(i.type_desc) + CASE 
        WHEN i.is_unique = 1
            THEN ', unique'
        ELSE ''
        END + CASE 
        WHEN i.is_primary_key = 1
            THEN ', primary key'
        ELSE ''
        END AS 'index_description'
    ,STUFF((
            SELECT ', [' + sc.NAME + ']' AS "text()"
            FROM syscolumns AS sc
            INNER JOIN sys.index_columns AS ic ON ic.object_id = sc.id
                AND ic.column_id = sc.colid
            WHERE sc.id = so.object_id
                AND ic.index_id = i1.indid
                AND ic.is_included_column = 0
            ORDER BY key_ordinal
            FOR XML PATH('')
            ), 1, 2, '') AS 'indexed_columns'
    ,STUFF((
            SELECT ', [' + sc.NAME + ']' AS "text()"
            FROM syscolumns AS sc
            INNER JOIN sys.index_columns AS ic ON ic.object_id = sc.id
                AND ic.column_id = sc.colid
            WHERE sc.id = so.object_id
                AND ic.index_id = i1.indid
                AND ic.is_included_column = 1
            FOR XML PATH('')
            ), 1, 2, '') AS 'included_columns'
FROM sysindexes AS i1
INNER JOIN sys.indexes AS i ON i.object_id = i1.id
    AND i.index_id = i1.indid
INNER JOIN sysobjects AS o ON o.id = i1.id
INNER JOIN sys.objects AS so ON so.object_id = o.id
    AND is_ms_shipped = 0
INNER JOIN sys.schemas AS s ON s.schema_id = so.schema_id
WHERE so.type = 'U'
    AND i1.indid < 255
    AND i1.STATUS & 64 = 0 --index with duplicates
    AND i1.STATUS & 8388608 = 0 --auto created index
    AND i1.STATUS & 16777216 = 0 --stats no recompute
    AND i.type_desc <> 'heap'
    AND so.NAME <> 'sysdiagrams'
ORDER BY table_name
    ,index_name;

- AeyJey

---

谢谢，这很有帮助。 - Sanket Sonavane

这太棒了！ - Unbreakable

很好。这个可以帮助您监视表中索引使用的列，方便您计划索引调优。 - YHTAN

sysobjects已被弃用... - SQLpro

数据库的观点如何？ - pRo

Answer 5

--简短明了：

SELECT OBJECT_SCHEMA_NAME(T.[object_id],DB_ID()) AS [Schema],  
  T.[name] AS [table_name], I.[name] AS [index_name], AC.[name] AS [column_name],  
  I.[type_desc], I.[is_unique], I.[data_space_id], I.[ignore_dup_key], I.[is_primary_key], 
  I.[is_unique_constraint], I.[fill_factor],    I.[is_padded], I.[is_disabled], I.[is_hypothetical], 
  I.[allow_row_locks], I.[allow_page_locks], IC.[is_descending_key], IC.[is_included_column] 
FROM sys.[tables] AS T  
  INNER JOIN sys.[indexes] I ON T.[object_id] = I.[object_id]  
  INNER JOIN sys.[index_columns] IC ON I.[object_id] = IC.[object_id] 
  INNER JOIN sys.[all_columns] AC ON T.[object_id] = AC.[object_id] AND IC.[column_id] = AC.[column_id] 
WHERE T.[is_ms_shipped] = 0 AND I.[type_desc] <> 'HEAP' 
ORDER BY T.[name], I.[index_id], IC.[key_ordinal]   

- Tim Ford

---

在连接到sys.index_columns时，您还需要加入I.index_id = IC.index_id。 - Chris J

FYI：上述查询在索引之间进行列的笛卡尔积。 - Saxman

Answer 6

我没有经历过，但我得到了原作者发布的查询中想要的结果。

我将其用于我的需求（没有条件/筛选），但它给出了错误的结果。

主要问题是结果在索引ID上没有连接条件而导致交叉乘积。

SELECT S.NAME SCHEMA_NAME,T.NAME TABLE_NAME,I.NAME INDEX_NAME,C.NAME COLUMN_NAME
  FROM SYS.TABLES T
       INNER JOIN SYS.SCHEMAS S
    ON T.SCHEMA_ID = S.SCHEMA_ID
       INNER JOIN SYS.INDEXES I
    ON I.OBJECT_ID = T.OBJECT_ID
       INNER JOIN SYS.INDEX_COLUMNS IC
    ON IC.OBJECT_ID = T.OBJECT_ID
       INNER JOIN SYS.COLUMNS C
    ON C.OBJECT_ID  = T.OBJECT_ID
   **AND IC.INDEX_ID    = I.INDEX_ID**
   AND IC.COLUMN_ID = C.COLUMN_ID
 WHERE 1=1

ORDER BY I.NAME,I.INDEX_ID,IC.KEY_ORDINAL

- dejjub-AIS

Answer 7

我需要获取特定索引及其索引列和包含列。这是我使用的查询：

SELECT INX.[name] AS [Index Name]
      ,TBL.[name] AS [Table Name]
      ,DS1.[IndexColumnsNames]
      ,DS2.[IncludedColumnsNames]
FROM [sys].[indexes] INX
INNER JOIN [sys].[tables] TBL
    ON INX.[object_id] = TBL.[object_id]
CROSS APPLY 
(
    SELECT STUFF
    (
        (
            SELECT ' [' + CLS.[name] + ']'
            FROM [sys].[index_columns] INXCLS
            INNER JOIN [sys].[columns] CLS 
                ON INXCLS.[object_id] = CLS.[object_id] 
                AND INXCLS.[column_id] = CLS.[column_id]
            WHERE INX.[object_id] = INXCLS.[object_id] 
                AND INX.[index_id] = INXCLS.[index_id]
                AND INXCLS.[is_included_column] = 0
            FOR XML PATH('')
        )
        ,1
        ,1
        ,''
    ) 
) DS1 ([IndexColumnsNames])
CROSS APPLY 
(
    SELECT STUFF
    (
        (
            SELECT ' [' + CLS.[name] + ']'
            FROM [sys].[index_columns] INXCLS
            INNER JOIN [sys].[columns] CLS 
                ON INXCLS.[object_id] = CLS.[object_id] 
                AND INXCLS.[column_id] = CLS.[column_id]
            WHERE INX.[object_id] = INXCLS.[object_id] 
                AND INX.[index_id] = INXCLS.[index_id]
                AND INXCLS.[is_included_column] = 1
            FOR XML PATH('')
        )
        ,1
        ,1
        ,''
    ) 
) DS2 ([IncludedColumnsNames])

- gotqn

---

我已经将我的原始创建语句与此查询的输出进行了比较。我可以说，对于我的数据库来说，这样正确地放置了索引和包含的列。 - bratak

这个脚本对我也起作用了。我有一个额外的要求，需要在另一个相同的数据库中重新创建索引，所以我将以下内容添加到SELECT语句中以生成CREATE语句：,CreateStatement = 'CREATE INDEX [' + INX.[name] + '] ON dbo.[' + TBL.[name] + '] (' + REPLACE(DS1.IndexColumnsNames, '] [', '], [') + ')' + CASE WHEN DS2.IncludedColumnsNames IS NOT NULL THEN ' INCLUDE (' + REPLACE(DS2.IncludedColumnsNames, '] [', '], [') + ')' ELSE '' END。 - Gary Pendlebury

Answer 8

以下内容类似于使用 sp_helpindex tablename

select T.name as TableName, I.name as IndexName, AC.Name as ColumnName, I.type_desc as IndexType 
from sys.tables as T inner join sys.indexes as I on T.[object_id] = I.[object_id] 
   inner join sys.index_columns as IC on IC.[object_id] = I.[object_id] and IC.[index_id] = I.[index_id] 
   inner join sys.all_columns as AC on IC.[object_id] = AC.[object_id] and IC.[column_id] = AC.[column_id] 
order by T.name, I.name

- hshen

Answer 9

基于接受的答案和其他两个问题^{1, 2}，我组装了以下查询：

SELECT
    QUOTENAME(t.name) AS TableName,
    QUOTENAME(i.name) AS IndexName,
    i.is_primary_key,
    i.is_unique,
    i.is_unique_constraint,
    STUFF(REPLACE(REPLACE((
        SELECT QUOTENAME(c.name) + CASE WHEN ic.is_descending_key = 1 THEN ' DESC' ELSE '' END AS [data()]
        FROM sys.index_columns AS ic
        INNER JOIN sys.columns AS c ON ic.object_id = c.object_id AND ic.column_id = c.column_id
        WHERE ic.object_id = i.object_id AND ic.index_id = i.index_id AND ic.is_included_column = 0
        ORDER BY ic.key_ordinal
        FOR XML PATH
    ), '<row>', ', '), '</row>', ''), 1, 2, '') AS KeyColumns,
    STUFF(REPLACE(REPLACE((
        SELECT QUOTENAME(c.name) AS [data()]
        FROM sys.index_columns AS ic
        INNER JOIN sys.columns AS c ON ic.object_id = c.object_id AND ic.column_id = c.column_id
        WHERE ic.object_id = i.object_id AND ic.index_id = i.index_id AND ic.is_included_column = 1
        ORDER BY ic.index_column_id
        FOR XML PATH
    ), '<row>', ', '), '</row>', ''), 1, 2, '') AS IncludedColumns,
    u.user_seeks,
    u.user_scans,
    u.user_lookups,
    u.user_updates
FROM sys.tables AS t
INNER JOIN sys.indexes AS i ON t.object_id = i.object_id
LEFT JOIN sys.dm_db_index_usage_stats AS u ON i.object_id = u.object_id AND i.index_id = u.index_id
WHERE t.is_ms_shipped = 0
AND i.type <> 0

此查询返回以下结果，显示索引列表、它们的列和使用情况。非常有帮助，可确定哪个索引比其他索引性能更好：

- Salman A

Answer 10

with connect(schema_name,table_name,index_name,index_column_id,column_name) as
(   select s.name schema_name, t.name table_name, i.name index_name, index_column_id, cast(c.name as varchar(max)) column_name
 from sys.tables t
inner join sys.schemas s on t.schema_id = s.schema_id
inner join sys.indexes i on i.object_id = t.object_id
inner join sys.index_columns ic on ic.object_id = t.object_id and ic.index_id=i.index_id
        inner join sys.columns c on c.object_id = t.object_id and
                ic.column_id = c.column_id
                where index_column_id=1
union all
select s.name schema_name, t.name table_name, i.name index_name, ic.index_column_id, cast(connect.column_name + ',' + c.name as varchar(max)) column_name
 from sys.tables t
inner join sys.schemas s on t.schema_id = s.schema_id
inner join sys.indexes i on i.object_id = t.object_id
inner join sys.index_columns ic on ic.object_id = t.object_id and ic.index_id=i.index_id
        inner join sys.columns c on c.object_id = t.object_id and
                ic.column_id = c.column_id join connect on
connect.index_column_id+1 = ic.index_column_id
and connect.schema_name = s.name
and connect.table_name = t.name
and connect.index_name = i.name)
select connect.schema_name,connect.table_name,connect.index_name,connect.column_name
from connect join (select schema_name,table_name,index_name,MAX(index_column_id) index_column_id
from connect group by schema_name,table_name,index_name) mx
on connect.schema_name = mx.schema_name
and connect.table_name = mx.table_name
and connect.index_name = mx.index_name
and connect.index_column_id = mx.index_column_id
order by 1,2,3

- jona

---

这对我很有帮助，因为我正在寻找表、索引名称和相应的列。 - Soman Dubey

据我所知，index_column_id 不是主键中列的顺序，而是表中的顺序！请使用 key_ordinal。 - Michel de Ruiter