Load order of tables in a CDC only task

[Ole_Dufour]

I'm wondering if setting the load order of tables in a CDC only task could change the data being replicated, or will it only impact performance ?

I have read the below documentation which I think only mentions performance :

https://help.qlik.com/en-US/replicate/May2024/Content/Global_Common/Content/SharedEMReplicate/Design...

[john_wang · Accepted solution]

Hello @Ole_Dufour ,

The table load order is relevant only during the Full Load stage. It does not apply to the CDC (Change Data Capture) stage. During CDC processing, change records are processed in the sequence they appear in the transaction log.

Hope this helps.

John.

[SushilKumar]

Hello @Ole_Dufour 

To add more to Expert comment. CDC mainly works Qr Scans for Transaction for participating's tables  in to change logs and once they are committed marked for replication to target . sometimes it depend upon the configured targets that they are applied on the target by the Sequence they initiated or commented on source.

Regards,

Sushil Kumar

[Heinvandenheuvel]

>>> change records are processed in the sequence they appear in the transaction log.

Yes and no. They are 'processed' in that order, but not 'applied' in that order.

Most CDC tasks are configure to run as "Batch Apply" notably when there is a performance concern as @Ole_Dufour  implies.

The order of changes defines which 'batch' a change ends up in.

But within the batch the Alphabetical table name defines which one gets applied first. In the end that does NOT affect performance since everything needs to be applied eventually.

Hein.

[Ole_Dufour]

Ok, merci Hein.