You can configure a custom mapping in the .NET configuration file to override the default mapping for the Number(p,0) Oracle data type. So, for example, Number(1,0), which is mapped to Int16 by default, can be custom mapped to the .NET Bool or .NET Byte type.
Example 4-1 shows a sample app.config file that uses custom mapping to map the Number(1, 0) Oracle data type to the bool EDM type. The example also maps Number(3,0) to byte, and sets the maximum precisions for the Int16, Int32, and Int64 data types to 4, 9, and 18 respectively.
Example 4-1 Sample Application Configuration File to Custom Map the Number (p,0) Data Type
<?xml version="1.0" encoding="utf-8"?>
<configuration>
<connectionStrings>
</connectionStrings>
<oracle.dataaccess.client>
<settings>
<add name="bool" value="edmmapping number(1,0)" />
<add name="byte" value="edmmapping number(3,0)" />
<add name="int16" value="edmmapping number(4,0)" />
<add name="int32" value="edmmapping number(9,0)" />
<add name="int64" value="edmmapping number(18,0)" />
</settings>
</oracle.dataaccess.client>
</configuration>
Example 4-2 shows a ODP.NET, Managed Driver sample app.config file.
Example 4-2 Sample ODP.NET, Managed Driver Application Configuration File to Custom Map the Number Data Type
<?xml version="1.0" encoding="utf-8" ?>
<configuration>
<oracle.manageddataaccess.client>
<version number="*">
<edmMappings>
<edmMapping dataType="number">
<add name="bool" precision="1"/>
<add name="byte" precision="3" />
<add name="int16" precision="4" />
<add name="int32" precision="9" />
<add name="int64" precision="18" />
</edmMapping>
</edmMappings>
</version>
</oracle.manageddataaccess.client>
</configuration>
Example 4-1 and Example 4-2 customizes the mappings as follows:
| Oracle Type | Default EDM Type | Custom EDM Type |
|---|---|---|
Number(1,0) |
Int16 |
bool |
Number(2,0) to Number(3,0) |
Int16 |
byte |
Number(4,0) |
Int16 |
Int16 |
Number(5,0) |
Int16 |
Int32 |
Number(6,0) to Number(9,0) |
Int32 |
Int32 |
Number(10,0) |
Int32 |
Int64 |
Number(11,0) to Number(18,0) |
Int64 |
Int64 |
Number(19,0) |
Int64 |
Decimal |
Custom mapping configures the maximum precision of the Oracle Number type that would map to the .NET/EDM type. So, for example, the preceding custom application configuration file configures ODP.NET to map Number(10,0) through Number(18,0) to Int64, as opposed to the default range of Number(11,0) through Number(19,0) for Int64.
Note:
Custom mapping does not require you to map all the .NET/EDM types. For example, if custom mapping is required just for Int16, then having a single entry for Int16 is sufficient. Default mapping gets used for the other types.
When using Model First, a Byte attribute is mapped to Number(3,0) by default. However, when a model is generated for a Number(3,0) column, it gets mapped to Int16 by default unless custom mapping for Byte is specified.
ODP.NET managed and unmanaged driver .NET configuration file settings for native Entity Framework 6 applications have a different format than the traditional .NET configuration file settings. For more information on this format, refer to the ODP.NET README. The README also describes ODP.NET-specific steps to migrate from an Entity Framework 5 application to Entity Framework 6.
You must make sure that your mappings allow the data to fit within the range of the .NET/EDM type and the Number(p, s) type. If you select a .NET/EDM type with a range too small for the Oracle Number data, then errors will occur during data retrieval. Also, if you select a .NET/EDM type, and the corresponding data is too big for the Oracle Number column, then INSERTs and UPDATEs to the Oracle database will error out.
To enable custom mapping, add the mapping information to the .NET config file prior to EDM creation.
If the EDM was created already before providing the mapping information, then you can modify the mappings either through the Visual Studio tools or manually. Using Visual Studio, go to the EDM Model Browser page. Right-click on the table(s) requiring new data type mapping and select Table Mapping from the pop-up menu. The Mapping Details window will appear usually at the bottom of your screen. Update Column Mappings as desired.
If you need to add or delete mappings, find the Type values in the CSDL mapping section of your project's existing EDMX file. Add or delete those Type values to the .NET data types you want the application to use. In the example below, the property name types for BOOLCOL and BYTECOL are added to the CSDL and mapped to Boolean and Byte, respectively.
Example Mapping Before CSDL Customization:
<Property Name="INT16COL" Type="Int16" Nullable="false" />
Example Mapping After CSDL Customization:
<Property Name="BOOLCOL" Type="Boolean" Nullable="false" /> <Property Name="BYTECOL" Type="Byte" Nullable="false" /> <Property Name="INT16COL" Type="Int16" Nullable="false" />
You can employ combinations of these customization possibilities depending on your planned mapping changes. If many tables and many columns require mapping changes, it is most efficient to delete the EDMX file and regenerate the data model. If a few tables and many columns require changes, then delete the affected tables, save the EDMX file, and select Update Model from Database... to include those tables again. If only a single table and one or two columns require changes, then modify the EDMX either manually or by using the Mapping Details window.
The following sections describe the Identity attribute and the Virtual column.
Oracle Database 12c (12.1) and later versions support table or view Identity attribute columns. Oracle has three Identity attribute types. When the EDM wizard generates a data model from an Oracle Identity attribute-containing table or view, ODP.NET will set the value of StoreGeneratedPattern to Identity in the .edmx file for any of three Oracle Identity types. The Identity attribute-associated column will use the server-generated value during INSERT: hence, application developers no longer need to create a sequence nor trigger. If the .NET application attempts to set the Identity attribute itself, this value will be ignored.
For Oracle Database 11g Release 2 (11.2) and earlier versions that do not support Identity columns, application developers can manually set StoreGeneratedPattern to Identity in columns through the entity model designer Properties after model generation, then create an INSERT trigger. Depending on the data type, a sequence may not be necessary if a server function, such as sys_guid(), can generate the value for the column.
Oracle Database 11g (11.1) and later versions can store expressions directly in base tables as Virtual columns, also known as Generated columns. Virtual columns cannot be inserted into or updated. ODP.NET will not automatically set StoreGeneratedPattern to Computed in the EF model for Virtual columns. To avoid errors, application developers need to add or change the value of StoreGeneratedPattern to Computed for Virtual columns after the model generation. Once done, Virtual columns are excluded from INSERTs and UPDATEs upon calling SaveChanges().
If the custom mapping in a .NET configuration file has changed, then regenerate the data model to solve compilation errors introduced by the changes.
Under certain scenarios, custom mapping may cause compilation errors when a project that uses custom mapping is loaded by Visual Studio. You may use the following workaround for such scenarios:
Open Visual Studio Help, About Microsoft Visual Studio. Click OK to exit the dialog box.
Alternatively, open the to-be-used connection in Server Explorer.
Compile the project again to eliminate the compilation errors.
When using your custom INSERT, UPDATE, or DELETE stored procedure in Stored Procedure Mapping, the following error might occur:
Error 2042: Parameter Mapping specified is not valid.
This can happen if a Number parameter has been mapped to a Boolean attribute, or if a RAW parameter has been mapped to a Guid attribute.
The solution is to manually add Precision="1" for the Number parameter, and MaxLength="16" for the RAW parameter of your stored procedure in the SSDL.