Attunity Replicate is a high-performance, easy-to-use data replication solution featuring industry-leading change data capture (CDC) technology. CDC eliminates batch processing by moving only the deltas to enable optimal efficiency in data transfer performance and real-time data availability.

Normally, Replicate scans a database’s transaction logs for changes and then applies those changes to the target database. However, this method of change processing is not possible with Data Warehouse endpoints such as Teradata Database or Microsoft APS PDW since these endpoints do not generate transaction logs.
The good news is that starting with Replicate 4.0.5 you can capture changes from Teradata or PDW databases - it just requires a little bit of preparation.

Before you can set a change data capture process from data warehouse source, you need to ensure that at least one special "Context" column exists in your source database tables. Context column(s) are basically columns in a table that enable Replicate to determine whether the data has changed. The values in the Context column has to be sequential, sorted in either ascending or descending order. You can add Context columns specifically for the purpose of change processing (either using a script or manually) or you can use existing columns that contain suitable "Context" data.

You can create and reference any number of Context columns in a table as long as the Context column names are the same for all source tables. Additionally, each value in the Context column(s) must be unique.
Name:  SampleRecords.png
Views: 1368
Size:  155.3 KB
In the example above, the 2 Context column, cf and oper, has been added to the table. The cf column contains TIMESTAMPs that enable Replicate to determine whether a change occurred (by comparing the current TIMESTAMP with the TIMESTAMP stored in its repository).

By default, all changes are assumed to be INSERTs. If UPDATE and DELETE operations are also performed on the source tables, the oper field can be used to capture type of operation, an UPDATE and/or DELETE, and will enable Replicate to identify the operation type.

All DELETE operations should be carried out as "Soft" deletes. This means that the row is not actually deleted from the table, but rather, marked as "deleted".

For details on how to configure a Change Data Capture process from a data warehouse source, please refer to Attunity Replicate Admin Guide.