Question

我有大约100个表，其中所有表都有重复的外键约束。

有没有办法可以摆脱它？是否有查询可以为我提供所有重复的密钥？

Answer 1

我在这里使用这个T-SQL脚本来检测可能重复的FK约束 - 它还在最后一个输出列中生成必要的ALTER TABLE...DROP CONSTRAINT语句。

您无法可靠地自动检测并选择要删除的多个FK约束中的哪一个 - 因此您基本上可以检测它们，然后手动选择要丢弃的那些（使用我的查询生成的drop语句）。

;WITH FKData AS
(
    SELECT  
        fk.parent_object_id,
        fkc.parent_column_id,
        fk.referenced_object_id,
        fkc.referenced_column_id,
        FKCount = COUNT(*)
    FROM 
         sys.foreign_keys fk
    INNER JOIN 
         sys.foreign_key_columns fkc ON fkc.constraint_object_id = fk.object_id
    GROUP BY 
         fk.parent_object_id, fkc.parent_column_id, fk.referenced_object_id, fkc.referenced_column_id
    HAVING 
         COUNT(*) > 1
),
DuplicateFK AS
(
    SELECT  
        FKName = fk.Name,
            ParentSchema = s1.Name,
        ParentTable = t1.Name,
        ParentColumn = c1.Name,
        ReferencedTable = t2.Name,
        ReferencedColumn = c2.Name 
    FROM 
        sys.foreign_keys fk
    INNER JOIN 
        sys.foreign_key_columns fkc ON fkc.constraint_object_id = fk.object_id
    INNER JOIN 
        FKData f ON fk.parent_object_id = f.parent_object_id 
                 AND fk.referenced_object_id = f.referenced_object_id 
                 AND fkc.parent_column_id = f.parent_column_id
                 AND fkc.referenced_column_id = f.referenced_column_id
    INNER JOIN 
        sys.tables t1 ON f.parent_object_id = t1.object_id
    INNER JOIN 
        sys.columns c1 ON f.parent_object_id = c1.object_id AND f.parent_column_id = c1.column_id
INNER JOIN 
    sys.schemas s1 ON t1.schema_id = s1.schema_id   
    INNER JOIN 
        sys.tables t2 ON f.referenced_object_id = t2.object_id
    INNER JOIN 
        sys.columns c2 ON f.referenced_object_id = c2.object_id AND f.referenced_column_id = c2.column_id
)
SELECT
    FKName,
    ParentSchema, ParentTable, ParentColumn, 
    ReferencedTable, ReferencedColumn,
    DropStmt = 'ALTER TABLE ' + ParentSchema + '.' + ParentTable + 
               ' DROP CONSTRAINT ' + FKName
FROM 
    DuplicateFK

Answer 2

对于100个表，它不是一个选项，但是如果您只有几个表，请在SQL Server Management Studio中创建一个图表，在那里添加表并直观地删除这些表。

Answer 3

这将删除最近创建的重复项

;WITH fkeys AS (
SELECT  f.object_id ,
        f.name ,
        f.parent_object_id,
        ROW_NUMBER() OVER(PARTITION BY t.column_names ORDER BY f.create_date,f.[object_id]) AS RowNum
FROM    sys.foreign_keys f
        CROSS APPLY ( SELECT    fc.parent_object_id,parent_column_id,fc.referenced_object_id ,fc.referenced_column_id
                      FROM      sys.foreign_key_columns fc
                      WHERE     fc.constraint_object_id = f.object_id
                      ORDER BY  constraint_column_id
                    FOR XML PATH('')
                    ) t ( column_names )
)
SELECT 'ALTER TABLE '+QUOTENAME(OBJECT_SCHEMA_NAME(f.parent_object_id)) + '.'+QUOTENAME(OBJECT_NAME(f.parent_object_id)) +' DROP CONSTRAINT '+QUOTENAME(f.name)+';' AS DropStatement
FROM fkeys f
WHERE f.RowNum >= 2

删除重复的外键

3 个答案: