DAC Support For SQL Server Objects and Versions
Applies To: SQL Server 2016 Preview
A data-tier application (DAC) supports the most commonly used Database Engine objects.
In This Topic
Only supported objects can be specified in a data-tier application as it is being authored or edited. You cannot extract, register, or import a DAC from an existing database that contains objects that are not supported in a DAC. SQL Server 2016 supports the following objects in a DAC.
FUNCTION: Inline Table-valued
FUNCTION: Multistatement Table-valued
STORED PROCEDURE: Transact-SQL
TABLE: Check Constraint
TABLE: Column, including computed columns
TABLE: Constraint, Default
TABLE: Constraint, Foreign Key
TABLE: Constraint, Index
TABLE: Constraint, Primary Key
TABLE: Constraint, Unique
TYPE: HIERARCHYID, GEOMETRY, GEOGRAPHY
TYPE: User-defined Data Type
TYPE: User-defined Table Type
The versions of SQL Server have different levels of support for DAC operations. All of the DAC operations supported by a version of SQL Server are supported by all editions of that version.
Instances of the Database Engine support the following DAC operations:
Export and extract are supported on all supported versions of SQL Server.
All operations are supported on Azure SQL Database and all versions of SQL Server 2014, SQL Server 2012, and SQL Server 2008 R2.
All operations are supported on SQL Server 2008 Service Pack 2 (SP2) or later, and SQL Server 2005 SP4 or later.
The DAC Framework comprises the client-side tools for building and processing DAC packages and export files. The following products include the DAC Framework
SQL Server 2014 and SQL Server 2012 includes DAC Framework 3.0, which supports all DAC operations.
SQL Server 2008 R2 SP1 and Visual Studio 2010 SP1 included DAC Framework 1.1, which supports all DAC operations except export and import.
SQL Server 2008 R2 and Visual Studio 2010 included DAC Framework 1.0, which supports all DAC operations except export, import, and in—place upgrade.
The client tools from earlier versions of SQL Server or Visual Studio do not support DAC operations.
A DAC package or export file built with one version of the DAC Framework cannot be processed by an earlier version of the DAC Framework. For example, a DAC package extracted using the SQL Server 2014 client tools cannot be deployed using the SQL Server 2008 R2 client tools.
A DAC package or export file built with one version of the DAC Framework can be processed by any later version of the DAC Framework. For example, a DAC package extracted using the SQL Server 2008 R2 client tools can be deployed using either the SQL Server 2008 R2 SP1 or higher client tools.
Note these fidelity limitations in the DAC Framework data deployment engine in SQL Server 2012 SP1. The limitations apply to the following DAC Framework actions: deploy or publish a .dacpac file, and import a .bacpac file.
Loss of metadata for certain conditions and base types within sql_variant columns. In the affected cases, you will see a warning with the following message: Certain properties on certain data types used within a sql_variant column are not preserved when deployed by the DAC Framework.
MONEY, SMALLMONEY, NUMERIC, DECIMAL base types: Precision is not preserved.
DECIMAL/NUMERIC base types with precision 38: the “TotalBytes” sql_variant metadata is always set to 21.
All text base types: The database default collation is applied for all text.
BINARY base types: Max length property is not preserved.
TIME, DATETIMEOFFSET base types: Precision is always set to 7.
Loss of data within sql_variant columns. In the affected case, you will see a warning with the following message: There will be data loss when a value in a sql_variant DATETIME2 column with scale greater than 3 is deployed by the DAC Framework. The DATETIME2 value is limited to a scale equal to 3 during deployment.
DATETIME2 base type with scale greater than 3: scale is limited to equal 3.
Deployment operation fails for the following conditions within sql_variant columns. In the affected cases, you will see a dialog with the following message: Operation failed due to data limitations in the DAC Framework.
DATETIME2, SMALLDATETIME and DATE base types: If the value is outside of DATETIME range – for example, the year is less than 1753.
DECIMAL, NUMERIC base type: when precision of the value is greater than 28.
Note the following considerations for DAC Framework data deployment actions:
Extract/Export - On actions that use the DAC Framework to create a package from a database – for example, extract a .dacpac file, export a .bacpac file - these limitations do not apply. The data in the package is a full-fidelity representation of the data in the source database. If any of these conditions are present in the package, the extract/export log will contain a summary of the issues via the messages noted above. This is to warn the user of potential data deployment issues with the package they created. The user will also see the following summary message in the log: These limitations do not affect the fidelity of the data types and values stored in the DAC package created by the DAC Framework; they only apply to the data types and values resulting from deploying a DAC package to a database. For more information about the data that is affected and how to work around this limitation, see this topic.
Deploy/Publish/Import - On actions that use the DAC Framework to deploy a package to a database, like to deploy or publish a .dacpac file, and import a .bacpac file, these limitations do apply. The data that results in the target database may not contain a full-fidelity representation of the data in the package. The Deploy/Import log will contain a message, noted above, for every instance the issue is encountered. The operation will be blocked by errors – see category 3 above - but will proceed with the other warnings.
For more information about the data that is affected in this scenario and how to work around this limitation for deploy/publish/import actions, see this topic.
Workarounds – Extract and export operations will write full-fidelity BCP data files into the .dacpac or .bacpac files. To avoid limitations, use the SQL Server BCP.exe command line utility to deploy full-fidelity data to a target database from a DAC package.