Naming, Non-historized Links and Source Data

Data Vault 2.0
1

I need some clarification on an aspect of non-historized links that I somehow failed to notice during my CDVP certification.

So, simply stated, links relate business concepts, and satellite provide the context. So far so simple. But part of providing said context, is that we split the satellites, as a minimum based on their source system origin. So one would have something like sat_account_erp and sat_account_crm.

Moving on to non-historized data, for example an accounting journal. As soon as we create a non-historized link, we combine the context with the relationship. And yet I have never seen or heard of a non-historized link being named after a source system, like nhlink_journal_erp1 and nhlink_journal_erp2. Why exactly?

If we model two journals from 2 source systems that link into the same business concepts (hub_gl_account, hub_cost_unit, …) shouldn’t the naming reflect the source origins?

1 Like
2

I think it should, good point!

3

NH links and sats are for real-time workloads that are immutable (non-historised) and therefore do not need to check the content being loaded is a true change.

Journal entries look like a batch entry, correct? If so then they should be loaded as satellites, or link-satellites.