I’m currently working on a Data Vault 2.0 project and have a question regarding modeling layers.
Do PIT and bridge tables belong to the Raw Vault or the Business Vault layer?
From my perspective, PIT tables serve purely as navigation aids and do not contain business logic. Similarly, bridge tables help with join paths and also do not implement business logic. In contrast, the Business Vault typically contains computed elements such as derived attributes, business rules, and reference tables (for integrating reference data).
Given this, I’m trying to understand why many sources place PIT and bridge tables in the Business Vault layer.
My general view is that PIT and bridge tables aren’t intrinsically tied to a single layer. Instead their placement should depend on what data they expose and how they’re used, rather than the table type itself.
Conceptually, I agree with your description:
PIT tables are primarily navigation/performance structures.
Bridge tables exist to simplify join paths (e.g. many‑to‑many or hierarchical relationships). On their own, neither must contain business logic.
Because of that, if a PIT or bridge:
is built purely from Raw Vault structures,
contains only raw keys and raw satellite references, and
has no derived attributes, filters, or business rules applied
then I don’t see any real harm in storing it alongside the Raw Vault. In that case it’s essentially a structural optimisation over raw history, not a business interpretation.
However, in practice, many PITs and bridges:
reference Business Vault satellites,
include derived or harmonised attributes,
apply business‑driven filters (e.g. “current”, “valid”, rule‑based status), or
are tailored to support downstream consumption patterns.
Once that happens, the object is no longer purely raw. At that point, placing it in the Business Vault (or even with marts) makes more sense, because it is now dependent on business rules and interpreted data.
I suspect this is why most guidance places PIT and bridge tables in the Business Vault: real-world implementations often mix raw and transformed data, and the BV becomes the safest and clearest home for them.
Hello @dylan.roe ,
sorry for the late reply and thank you very much for such a full exlplanation. I had to restructure my model and move pit and bridges tablöes to business vault layer exactly because of some filters included in its logic.
