The Disconnected Source Model Load feature of SQL Developer allows consultants to work on a customer's database migration without having to install and run SQL Developer at the customer site.
To perform the disconnected source model load option a customer must generate delimited flat files containing schema metadata from the database to be migrated. You generate the flat file by running a predefined SQL Developer script against the source database. The flat files are sent to a consultant who uses SQL Developer to load the metadata files into a source and Oracle model. You can then map this schema to Oracle.
Microsoft SQL Server or Sybase Adaptive Server databases use the Bulk Copy Program (BCP) to generate delimited metadata flat files. Predefined scripts installed with SQL Developer invoke the BCP, and generate the flat files for each database. The BCP outputs delimited metadata files from the database with a .dat extension. However, for a successful migration of a database the .dat metadata files are converted into XML files by SQL Developer. SQL Developer converts the .dat files when the source metadata files are selected during the capture phase of the migration, and outputs the generated .xml files to the same root directory as the source .dat files.
The predefined script files are stored in the %ORACLE_HOME%\Omwb\DSML_Scripts\plugin directory. Refer to the following table to locate the correct SQL Developer script:
Table 5-1 Location and Name of Script Files and Name of Associated Files
| Plug-in | Directory Location | Script File Name | Associated Files |
|---|---|---|---|
|
Microsoft SQL Server 7 |
|
SS7_DSML_SCRIPT.BAT |
Not Applicable |
|
Microsoft SQL Server 2000 |
|
SS2K_DSML_SCRIPT.BAT |
Not Applicable |
To run a script file for a plugin from the %ORACLE_HOME%\DSML_scripts\<plugin> directory, use the following command line:
<script file name> <database> <password> <server>
For example, to run the Microsoft SQL Server 2000 script file to generate metadata flat files, use the following command:
SS2K_DSML_SCRIPT <database> <password> <server>