I executed below scenarios and the behaviour seems to be inconsistent. I noticed cdc.change_tables tracks table details with index_name but BOL doesn't explain what is/isn't possible interms of modifying PK. Is #1 by design, if so it needs to be clarified.
- Enable CDC on a table with PK. Later try to disable/drop/change PK definition on base table – It is not allowed
- Enable CDC on a table with no PK. Later try to create/change/drop PK definition on base table – It is allowed
Note: NET changes is not enabled in both cases.
Thanks,
Siva
let me get back to you on this...|||
I see you've already filed this issue in connect. I'll cut/paste the response here as well for anyone else that's wondering the same thing.
The behavior is by design. When CDC is enabled and if a primary key exists on the table, CDC will use the index regardless of whether net changes is enabled or not.
If there is no primary key on the table, you can still enable CDC but only with net changes set to false. You are then able to create a primary key and alter it since CDC does not use the PK.
This will be documented in BOL.
No comments:
Post a Comment