On the auditing side of things, we want to see when items are deleted.
I read Cuba’s “Data Vault’s New Hero” post about RTS and XTS.
Do I understand correctly that in an RTS, you’re recording the state on every load?
In a CDC system, I can see this. You’re only tracking changes. But in a snapshot system, with millions of rows in a table being loaded every day, how would an RTS work?
Cuba’s book has an example of an STS that captures inserts and deletes, which totally works and I am prototyping now, so our audit may be solved there, but it just has me wondering if an RTS is completely off the menu for snapshot systems.
Thanks,
Joe
1 Like
Hi Joe, welcome to the forum.
An RTS do record every state on every load. But that does not mean that you can not load it in a slim way. You can load the RTS as a delta, a delta of the state. This means that if a dataset includes businees key X in 100 consecutive loads only one row is inserted into the RTS. If load number 101 does not contain businees key X than the state has changed and another row must be inserted into the RTS. I guess this is what you are looking for.
RTS only records that the record existed — not its state
XTS is used for time-line correction
STS (status tracking) tracks the status, following on from your previous post your source is a snapshot every time and therefore an easy candidate for an STS if the source does not supply a status update