Question

我有一个现有的数据库（SSDT项目作为源代码），我必须在其中更新表格。我只是将列类型从nvarchar(MAX)更改为nvachar(256)。问题是，我有一个与SCHEMABINDING绑定到此表的现有视图。我无法编辑表格列，因为SCHEMABINDING会阻止影响VIEW的更改。

以下脚本在PreDeploymentScript中执行。当我调用DROP VIEW [base].[VIEW_DEPENDING_ON_TABLE]语句时，部署后缺少视图。我的想法是，在部署期间禁用SCHEMABINDING并在完成后启用它。在TSQL脚本中这怎么可能？或者有更好的方法吗？

IF EXISTS ( SELECT DATA_TYPE FROM INFORMATION_SCHEMA.COLUMNS
        WHERE TABLE_SCHEMA = N'base'
        AND TABLE_NAME = 'TABLENAME'
        AND COLUMN_NAME = 'Instance'
        AND CHARACTER_MAXIMUM_LENGTH = -1)
AND NOT EXISTS (SELECT * FROM sys.indexes WHERE name='IX_TABLENAME_Instance' AND object_id = OBJECT_ID(N'[base].[TABLENAME]'))
BEGIN

IF EXISTS (SELECT * FROM sys.objects WHERE object_id = OBJECT_ID(N'[base].[VIEW_DEPENDING_ON_TABLE]') AND type in (N'V'))
    BEGIN
        DROP VIEW [base].[VIEW_DEPENDING_ON_TABLE]
    END

IF EXISTS (SELECT * FROM sys.objects WHERE object_id = OBJECT_ID(N'[base].[TABLENAME]') AND type in (N'U'))
    BEGIN
        UPDATE [base].[TABLENAME] SET [Instance] = LEFT(Instance, 256)
        ALTER TABLE [base].[TABLENAME] ALTER COLUMN [Instance] NVARCHAR(256)
    END

IF EXISTS (SELECT * FROM sys.objects WHERE object_id = OBJECT_ID(N'[base].[TABLENAME]') AND type in (N'U'))
    BEGIN
        CREATE NONCLUSTERED INDEX [IX_TABLENAME_Instance] ON [base].[TABLENAME]
        (
            [Instance] ASC
        ) WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, SORT_IN_TEMPDB = OFF, DROP_EXISTING = OFF, ONLINE = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
    END
END
GO

Answer 1

没有命令禁用WITH SCHEMABINDING。您可以通过删除视图并在没有选项的情况下重新创建视图，或者只是更改它来执行此操作。在你的情况下：

ALTER VIEW [base].[VIEW_DEPENDING_ON_TABLE]
-- WITH SCHEMABINDING
AS
    SELECT ...;

对基础表进行更改后，可以再次更改视图：

ALTER VIEW [base].[VIEW_DEPENDING_ON_TABLE]
WITH SCHEMABINDING
AS
    SELECT ...;

没有方便的脚本？没关系，您可以从sys.sql_modules提取，但需要进行一些按摩才能从CREATE转换为ALTER并删除WITH SCHEMABINDING。您可以编写解析它的代码，但这非常脆弱，因为替换或评论短语“with schemabinding”可能非常困难 - 它可能有多个空格，制表符，回车符，非打印字符等，它也可能存在于代码的其他部分（如注释，甚至是表或列别名）。

更改现有VIEW的SCHEMABINDING

1 个答案: