Define table and column mappings - Connect_CDC - connect_cdc_mimix_share - Latest

Once you have made the source and target tables available to the servers in the model (using the Refresh Available Tables menu item), you define to the Connect CDC Director which target table(s) receive data from the source table.

1. Select a table for distribution from the first server.

   • From the list of available tables on the first server that has source tables, select Select for Distribution >This Table Only.

   For example, select DIVISION1.PAYROLL on server ORCL.

   The Distribute Tables dialog appears.

   The name of the table you are distributing is in the top box.

   • Select the server to which you are distributing the source table.

   • Select a Distribution from the list of available distributions if any have been defined.

     • Click OK to close the Distribute Tables dialog.

2. On the Sending Table Properties dialog box, Table tab for the source table, this page has a default for the bi-directional setup:

   Full resolution, by shadow table is necessary for bidirectional replication and is checked under Collision detection and resolution during model validation. It is recommended that you accept the generated shadow table name.

   A shadow table is used to detect when two different replication databases have pending updates against the same row from a table. If this situation is not detected, the data will be out of sync at each database when the updates are applied at the targets. The shadow table allows this situation to be detected and resolved so that the data is consistent at both source databases.

   Each replicated table will have a corresponding shadow table. When a replicated source table is modified, if there is no existing entry, a new entry is added by the triggers to the shadow table for each unique row . If there is an existing entry, the triggers update the entry with the timestamp, transaction ID and additional information that is used to detect collisions.

   When the kernel is applying data at a target, it queries the shadow table using the primary key values of the current row. The information from the shadow row is compared to the current row to be applied to determine if there was a conflicting update.

3. If source tables on the two participating servers have different names, map the source table on the first server to the source table on the second server using the Sending Table Properties, Mapping tab.

   For example, map DIVISION2.PAYROLL2 to receiving table DIVISION1.PAYROLL.

4. Click OK to close the Sending Table Properties.

5. Select a table for distribution from the second server.

   1. From the list of available tables for walt-patils2 that holds the second source table, select for distribution the table you mapped to the first source table.

      For example, select PATILS.LONGCHAR1. The Distribute Tables dialog appears.

      The name of the table you are distributing displays in the top box.

   2. Select the server to which you want to distribute the table.

   3. Click OK.

6. If source tables on the two participating source servers have different names, map the source table on second server to the source table on the first server using the Sending Table Properties, Mapping tab.

   For example, map Source.dbo.Employee_Details on walt-patils2 to PATILS.LONGCHAR1 on orcl.

7. Click OK to close the Sending Table Properties.