Enabling Change Tracking on Target(Snowflake) tables

[adapavandana]

What if I enabling CHANGE_TRACKING on the target table, does Qlik continue to refresh the table?  is there a spike in costs? My target is Snowflake and source is IBM i.

[OritA · Accepted solution]

Hi, 

I am not sure that I fully understand the scenarion but, in genral: 
When you Replicate from DB2i to any other target, then the CDC changes are read by Replicate from the source endpoint and written to the target endpoint. Any action that happens after the  event  is written to the target does not affect Replicate itself.  It might have a side affect on the resource consumption on the machine and as result affect the processes that are working in this environment at that time. 

Regards,

Orit

[adapavandana]

Hi Orit,

Here the need is , must know all INSERTS/UPDATES/DELETES from a given table during a given timeframe.  Including on tables that are lacking accurate LastModifiedDate timestamps for each of these scenarios. 

* Key here is we also need to know about any deletes since the last time we checked.

[Heinvandenheuvel]

You can count delete by changing them to 'soft' delete by setting a flag in a new column for example called 'is_deleted' or 'row_status'. Within Replicate you would add the  column with an expression to load a value using the OPERATION_INDICATOR function.

Alternatively you may choose to enable the STORE_CHANGES task option where Replicate will create a changes table (often named xxx_ct) per table.

Good luck,

Hein.

[Michael_Litz]

Hi adapavandana,

Yes, the change table will record each transaction for the table in a separate table with the __ct suffix. 

With the store changes enabled the task will still update the base table to keep it in sync.

There is a little more resource overhead as the task will need to write to this __ct table for each transaction in addition to update of the base table.

Thanks,
Michael