If you add or drop columns in a Microsoft SQL Server source table, you must create a new capture registration. PowerExchange generates an associated extraction map by default. Use the generated extraction map instead of creating a user-defined extraction map. The PWX CDC Reader on the MySQL CDC Connector system will retrieve information from the new generated extraction map.
To switch to the new table definition and preserve access to previously captured data, you must perform several steps in both the PowerExchange and SQL Server CDC Connector environments.
You do not need to perform these steps if you selectively capture change data for a subset of columns, and the DDL changes do not affect any of these columns.
Stop data change activity on the SQL Server source table.
Verify that any change data that was captured under the current table definition has completed SQL Server CDC Connector extraction processing. Then stop all mapping tasks that extract change data for the table.
Shut down the PowerExchange Logger for Linux, UNIX, and Windows.
In the PowerExchange Navigator, open the original capture registration and set its status to
History
.
Add or drop columns in the source table.
In the PowerExchange Navigator, create a new capture registration for the table that reflects the column changes. Include the following settings:
In the
Condense
list, select
Part
.
In the
Status
list, select
Active
.
Warm start the PowerExchange Logger.
Alter the target table to reflect the source table changes, if necessary.
In Data Integration, for each mapping that includes the source table, perform the following steps:
Select the Source transformation. Under Source Properties, either click the
Synchronize
button on the
Fields
tab or select the table again in the
Object
field on the
Source
tab. Either action causes the PWX CDC Metadata Adapter to retrieve the updated extraction-map metadata.
If you changed the target table, select the Target transformation. Under Target Properties, click the
Synchronize
button on the
Target Fields
tab or select the table again in the
Object
field on the
Target
tab.
Edit the field mapping to reflect the added or removed columns.
If necessary, rematerialize the target tables.
In the mapping task, set a time-based restart point that coincides with the time of target rematerialization or let the mapping task restart from the last checkpoint that was recorded in the state table on the target.