LEFT JOIN sys.default_constraints dc WITH(NOLOCK) ON dc. = c.default_object_id LEFT JOIN sys.identity_columns ic WITH(NOLOCK) ON LEFT JOIN sys.computed_columns cc WITH(NOLOCK) ON JOIN sys.schemas s WITH(NOLOCK) ON tp. = s. JOIN sys.types tp WITH(NOLOCK) ON c.user_type_id = tp.user_type_id This forces joins to other system views to be established as in the following: SELECT However, data contained in sys.columns (as well as in INFORMATION_SCHEMA.COLUMNS) is not enough to completely describe the table structure. In terms of performance, the #4 approach remains the most attractive to me. The presented plans show that the #1 and #2 approaches contain excessive amount of connections that will increase the query execution time, while the #3 approach leads to the complete scan of the index, making it the least efficient of all. TimeSheetDate AS DATEADD(DAY, -(DAY(DateOut) - 1), DateOut),ĬONSTRAINT PK_WorkOut PRIMARY KEY CLUSTERED (WorkOutID) To start, let's create a test table, in order for the process of script writing to be more clear: IF OBJECT_ID('dbo.WorkOut', 'U') IS NOT NULL

Well, let’s look at system views and create an OBJECT_DEFINITION function analogue for working with table objects. Sometimes it is neecessary to retrieve a script description of a table for some scripts. Fetching from sys.sql_modules will also not solve the problem, since the same old OBJECT_DEFINITION function call is used inside this system view:

Mysql create table from existing table