Question

在我相当大的数据库中，我想知道在整个模式（SP，函数，表，触发器......）中引用列的所有地方。我不想只进行文本搜索，因为这将获取注释，并且还会从其他表中找到类似命名的列。

有谁知道我是否可以这样做？我使用的是SQL Server 2008。

Answer 1

警告：即使这是一种文本搜索方法，我要分享的脚本也为我节省了很多很多时间。它在里面搜索：

标量函数
表值函数
存储过程
视图
触发器

我需要指定排序规则以使其适用于我。

SELECT sys.objects.object_id, sys.schemas.name AS [Schema], sys.objects.name AS Object_Name, sys.objects.type_desc AS [Type]
FROM sys.sql_modules (NOLOCK) 
INNER JOIN sys.objects (NOLOCK) ON sys.sql_modules.object_id = sys.objects.object_id 
INNER JOIN sys.schemas (NOLOCK) ON sys.objects.schema_id = sys.schemas.schema_id
WHERE
    sys.sql_modules.definition COLLATE SQL_Latin1_General_CP1_CI_AS LIKE '%{Column Name}%' ESCAPE '\'
ORDER BY sys.objects.type_desc, sys.schemas.name, sys.objects.name

输出如下：

Output

更新：如果您需要搜索某个表格，SP等，您可以使用更专业的查询：

DECLARE @SCHEMA_NAME VARCHAR(100) = 'dbo';
DECLARE @OBJECT_NAME  VARCHAR(100) = 'MY_OBJECT';

SELECT
    sys.objects.object_id,
    sys.schemas.name AS [Schema], 
    sys.objects.name AS Object_Name, 
    sys.objects.type_desc AS [Type]
FROM sys.sql_modules (NOLOCK) 
INNER JOIN sys.objects (NOLOCK) ON sys.sql_modules.object_id = sys.objects.object_id 
INNER JOIN sys.schemas (NOLOCK) ON sys.objects.schema_id = sys.schemas.schema_id
WHERE
    (
           '#' + sys.sql_modules.definition + '#' COLLATE SQL_Latin1_General_CP1_CI_AS LIKE '%[^a-z_]'+@SCHEMA_NAME+'.'+@OBJECT_NAME+'[^a-z_]%' ESCAPE '\'
        OR '#' + sys.sql_modules.definition + '#' COLLATE SQL_Latin1_General_CP1_CI_AS LIKE '%[^a-z_]\['+@SCHEMA_NAME+'\].'+@OBJECT_NAME+'[^a-z_]%' ESCAPE '\'
        OR '#' + sys.sql_modules.definition + '#' COLLATE SQL_Latin1_General_CP1_CI_AS LIKE '%[^a-z_]'+@SCHEMA_NAME+'.\['+@OBJECT_NAME+'\][^a-z_]%' ESCAPE '\'
        OR '#' + sys.sql_modules.definition + '#' COLLATE SQL_Latin1_General_CP1_CI_AS LIKE '%[^a-z_]\['+@SCHEMA_NAME+'\].\['+@OBJECT_NAME+'\][^a-z_]%' ESCAPE '\'
    )
ORDER BY sys.objects.type_desc, sys.schemas.name, sys.objects.name

P.S。：两个查询也会在评论中搜索。

Answer 2

this article中描述了执行此操作的最佳方式。

样本：

SELECT OBJECT_NAME (referencing_id),
              referenced_database_name, 
       referenced_schema_name, referenced_entity_name
FROM sys.sql_expression_dependencies d
WHERE OBJECT_NAME(d.referenced_id) = 'TableName'
      AND OBJECT_DEFINITION (referencing_id)  LIKE '%ColumnName%'
ORDER BY OBJECT_NAME(referencing_id);

Answer 3

我尝试了这个查询，似乎很好：

select 
obj.type REFERENCING_OBJECT_TYPE
 ,SCHEMA_NAME(obj.schema_id) REFERENCING_OBJECT_SCHEMA
 ,obj.name                  REFERENCING_OBJECT_NAME
from sysdepends x
INNER JOIN sys.objects obj ON x.id  = obj.object_id
where depid = object_id('yourSchema.yourTable')
and col_name(depid, depnumber) = 'yourColumn'

Answer 4

正如Luv所说，这是一个老问题，但我发现了另外两个可能有用的解决方案。

我使用sys.dm_sql_referenced_entities系统对象查找指定对象中所有引用的对象和列。您可以使用以下查询：

SELECT DISTINCT 
referenced_schema_name AS SchemaName, 
referenced_entity_name AS TableName, 
referenced_minor_name  AS ColumnName
FROM sys.dm_sql_referenced_entities ('yourrefencingobject', 'OBJECT'); 
GO

这给出了以下结果：

enter image description here

此对象的缺点是您需要指定引用对象。

或者进行搜索：

SELECT DISTINCT object_name(id) 
FROM AdventureWorks2012.dbo.syscomments (nolock) 
WHERE text like '%BusinessEntityID%'

这给出了以下结果：

enter image description here

我还发现了您可以在this article中使用的以下SP，但尚未对其进行正确测试：

> DECLARE    @string varchar(1000),   @ShowReferences char(1)
> 
> SET @string = 'Person.Person.BusinessEntityID' --> searchstring
> 
> SET @ShowReferences = 'N'
> /****************************************************************************/ /*                                                                    
> */ /* TITLE:   sp_FindReferences                                               */ /*                                                                          */ /* DATE:    18 February, 2004                                               */ /*                                                                          */ /* AUTHOR:  WILLIAM MCEVOY                                                  */ /*                                                                          */ /****************************************************************************/ /*                                                                    
> */ /* DESCRIPTION:  SEARCH SYSCOMMENTS FOR INPUT STRING, OUTPUT NAME OF OBJECT */ /*                                                       
> */ /****************************************************************************/ set nocount on
> 
> declare @errnum         int         ,
>         @errors         char(1)     ,
>         @rowcnt         int         ,
>         @output         varchar(255)
> 
> select  @errnum         = 0         ,
>         @errors         = 'N'       ,
>         @rowcnt         = 0         ,
>         @output         = ''        
> 
> /****************************************************************************/ /* INPUT DATA VALIDATION                                              
> */ /****************************************************************************/
> 
> 
> /****************************************************************************/ /* M A I N   P R O C E S S I N G                                      
> */ /****************************************************************************/
> 
> -- Create temp table to hold results DECLARE @Results table (   Name        varchar(55),   Type        varchar(12),   DateCreated datetime,  
> ProcLine    varchar(4000) )
> 
> 
> IF (@ShowReferences = 'N') BEGIN   insert into @Results   select
> distinct
>          'Name' = convert(varchar(55),SO.name),
>          'Type' = SO.type,
>          crdate,
>          ''
>     from sysobjects  SO
>     join syscomments SC on SC.id = SO.id    where SC.text like '%' + @string + '%'   union   select distinct
>          'Name' = convert(varchar(55),SO.name),
>          'Type' = SO.type,
>          crdate,
>          ''
>     from sysobjects  SO    where SO.name like '%' + @string + '%'   union   select distinct
>          'Name' = convert(varchar(55),SO.name),
>          'Type' = SO.type,
>          crdate,
>          ''
>     from sysobjects  SO
>     join syscolumns SC on SC.id = SO.ID    where SC.name like '%' + @string + '%'    order by 2,1 END ELSE BEGIN   insert into @Results  
> select 
>          'Name'      = convert(varchar(55),SO.name),
>          'Type'      = SO.type,
>          crdate,
>          'Proc Line' = text
>     from sysobjects  SO
>     join syscomments SC on SC.id = SO.id    where SC.text like '%' + @string + '%'   union   select 
>          'Name'      = convert(varchar(55),SO.name),
>          'Type'      = SO.type,
>          crdate,
>          'Proc Line' = ''
>     from sysobjects  SO    where SO.name like '%' + @string + '%'   union   select 
>          'Name' = convert(varchar(55),SO.name),
>          'Type' = SO.type,
>          crdate,
>          'Proc Line' = ''
>     from sysobjects  SO
>     join syscolumns SC on SC.id = SO.ID    where SC.name like '%' + @string + '%'    order by 2,1 END
> 
> IF (@ShowReferences = 'N') BEGIN   select Name,
>          'Type' = Case (Type)
>                     when 'P'  then 'Procedure'
>                     when 'TR' then 'Trigger'
>                     when 'X'  then 'Xtended Proc'
>                     when 'U'  then 'Table'
>                     when 'C'  then 'Check Constraint'
>                     when 'D'  then 'Default'
>                     when 'F'  then 'Foreign Key'
>                     when 'K'  then 'Primary Key'
>                     when 'V'  then 'View'
>                     else Type
>                   end,
>          DateCreated
>     from @Results
>     order by 2,1 END ELSE BEGIN   select Name,
>          'Type' = Case (Type)
>                     when 'P'  then 'Procedure'
>                     when 'TR' then 'Trigger'
>                     when 'X'  then 'Xtended Proc'
>                     when 'U'  then 'Table'
>                     when 'C'  then 'Check Constraint'
>                     when 'D'  then 'Default'
>                     when 'F'  then 'Foreign Key'
>                     when 'K'  then 'Primary Key'
>                     when 'V'  then 'View'
>                     else Type
>                   end,
>          DateCreated,
>          ProcLine
>     from @Results
>     order by 2,1 END

希望这有帮助

Answer 5

您好虽然这是一篇旧帖子，但我能够将上面提到的一些提示结合起来，这对我来说很有帮助。我创建这个的原因是我在很多表中的列，所以它没有提供清晰的输出。

SELECT 
      SCHEMA_NAME(schema_id)+'.'+[name] as objectname
      ,type_desc
     ,referenced_schema_name AS SchemaName
,referenced_entity_name AS TableName
,referenced_minor_name  AS ColumnName
  FROM [sys].[all_objects] ob cross apply sys.dm_sql_referenced_entities ( SCHEMA_NAME(schema_id)+'.'+[name], 'OBJECT') e

  where is_ms_shipped = 0 and type_desc in ('AGGREGATE_FUNCTION'
,'SQL_SCALAR_FUNCTION'
,'SQL_INLINE_TABLE_VALUED_FUNCTION'
,'SQL_STORED_PROCEDURE'
,'SQL_TABLE_VALUED_FUNCTION'
,'SQL_TRIGGER'
,'VIEW')
and name !='sp_upgraddiagrams'
and referenced_entity_name  = 'table name'
and referenced_minor_name = 'columnname'

Answer 6

我想避免使用sys.sql_dependencies，因为将在Microsoft SQL Server的未来版本中删除此功能。我也不能使用OBJECT_DEFINITION函数，因为我的所有对象都是加密的。所以我提出了以下查询，这看起来非常简单，对我来说非常有用，我甚至把它包装到函数中：

-- =============================================
-- Description: Gets all the stored procedures, functions and triggers referencing the specified column.
-- Example:     SELECT * FROM dbo.UFN_GET_SP_FN_TR_REFERENCING_COLUMN(N'dbo', N'MY_TABLE', N'MY_COLUMN');
-- =============================================
CREATE FUNCTION dbo.UFN_GET_SP_FN_TR_REFERENCING_COLUMN 
(
    @SchemaName sysname,
    @TableName sysname,
    @ColumnName sysname
)
RETURNS TABLE
AS 
RETURN

    SELECT R.referencing_schema_name + N'.' + R.referencing_entity_name AS referencing_entity_name
    FROM sys.dm_sql_referencing_entities(@SchemaName + N'.' + @TableName, 'OBJECT') AS R 
        INNER JOIN sys.objects AS O 
            ON R.referencing_id = O.object_id
    WHERE O.[type] IN ('FN'/*SQL scalar function*/, 'IF'/*SQL inline table-valued function*/, 'TF'/*SQL table-valued-function*/, 'P'/*SQL Stored Procedure*/, 'TR' /*SQL DML trigger*/)
          AND EXISTS(SELECT 1 FROM sys.dm_sql_referenced_entities (R.referencing_schema_name + N'.' + R.referencing_entity_name, 'OBJECT') AS RE WHERE RE.referenced_entity_name = @TableName AND RE.referenced_minor_name = @ColumnName);

GO

Answer 7

这里对@ alex的TV-UDF稍作调整，以包含观点：

/*
    Source: https://stackoverflow.com/a/47775531/852956
    Gets all the stored procedures, functions and triggers referencing the specified column.

    SELECT * FROM Utility.ft_SelectSprocFuncAndTrigrRefs(N'BrakeRotor', N'BrakeRotors', N'BrakeRotorNumber');
*/
CREATE FUNCTION Utility.ft_SelectSprocFuncAndTrigrRefs
(
    @SchemaName sysname,
    @TableName sysname,
    @ColumnName sysname
)
RETURNS TABLE
AS 
RETURN

    SELECT QUOTENAME(R.referencing_schema_name) + N'.' + QUOTENAME(R.referencing_entity_name) AS ReferencingEntityName
    FROM sys.dm_sql_referencing_entities(@SchemaName + N'.' + @TableName, 'OBJECT') AS R 
        INNER JOIN sys.objects AS O 
            ON R.referencing_id = O.object_id
    WHERE O.[type] IN (
                'FN'/*SQL scalar function*/,
                'IF'/*SQL inline table-valued function*/,
                'TF'/*SQL table-valued-function*/,
                'P'/*SQL Stored Procedure*/,
                'TR' /*SQL DML trigger*/
            )
          AND EXISTS(
                    SELECT 1
                    FROM sys.dm_sql_referenced_entities (R.referencing_schema_name + N'.' + R.referencing_entity_name, 'OBJECT') AS RE
                    WHERE RE.referenced_entity_name = @TableName AND RE.referenced_minor_name = @ColumnName)

    UNION SELECT QUOTENAME(VIEW_SCHEMA) + N'.' + QUOTENAME(VIEW_NAME) AS ReferencingEntityName
        FROM INFORMATION_SCHEMA.VIEW_COLUMN_USAGE
        WHERE TABLE_SCHEMA = @SchemaName
          AND TABLE_NAME = @TableName
          AND column_name = @ColumnName
GO

Answer 8

今天不得不这样做并想出了以下内容：

import matplotlib.pyplot as plt
import pandas as pd
import matplotlib.ticker as ticker
%matplotlib inline

plt.figure(figsize=(15,6))

ax1 = plt.subplot(2,5,1)
plt.plot(tickers_data.index.strftime("%d"),tickers_data['Volume','XPEV'])
plt.ticklabel_format(style='plain', axis='y')
plt.title('XPEV')

ax2 = plt.subplot(2,5,2)
plt.plot(tickers_data.index.strftime("%d"),tickers_data['Volume','M'])
plt.ticklabel_format(style='plain', axis='y')
plt.title('M')

ax3 = plt.subplot(2,5,3)
plt.plot(tickers_data.index.strftime("%d"),tickers_data['Volume','MLCO'])
plt.ticklabel_format(style='plain', axis='y')
plt.title('MLCO')

ax4 = plt.subplot(2,5,4)
plt.plot(tickers_data.index.strftime("%d"),tickers_data['Volume','VIPS'])
plt.ticklabel_format(style='plain', axis='y')
plt.title('VIPS')

ax5 = plt.subplot(2,5,5)
plt.plot(tickers_data.index.strftime("%d"),tickers_data['Volume','HD'])
plt.ticklabel_format(style='plain', axis='y')
plt.title('HD')

在SQL Server中，如何在任何位置引用列？

8 个答案: