Hub with many multi-active satellites

Hi. I have a raw data vault model that has a hub with about 15 satellites (all from a single source system) hanging off it. About half of those satellites are multi-active. My question is, is this unusual or is this common?

My first thought is that there are some hubs that have been missed, and that we need to dig a bit deeper to see if we can discover more hubs.

Thanks in advance.

All from the same source system?

Is this unusual? Probably yeah.
Is it wrong? No.
Is it efficient? Likely not.

If there’s a good use case for splitting the data like this then I can imagine it being feasible. PII, Legal requirements, grain differences, update frequency etc.

How many columns are being split here? If you’ve got thousands of attributes here I can understand the segmentation of what is in each satellite but if these are slim tables then I don’t know if I’d approach it the same way you have haha.

All the best
Frankie

I had Tom Jones in my head.

Well, I was waiting for someone to answer before I put my 2c in.

  • Why not?
  • Are there problems with getting the data out?
  • Is the data appropriately split?
  • What was the use case for using MSATS? Are they true MSATs or should they be Satellites with dependent-child keys?
  • Why would you need more hub tables? Are the representative of the Business Architecture or are you just finding codes in satellites and thinking maybe that should be a hub? Maybe they are dependent-child keys?

MSATs are rare, dependent-child keys should occur more often for multi-row tracking to a business object.

Just to be clear, the source for each satellite is a different source system table. It is not one huge table that is split into different satellites. Also, the majority of the are multi-active.

I did not get that he is splitting the source tables into multiple satellites, just that there are many tables from a single source system that all roll up to a single entity/hub.

Without knowing what the tables are and their business keys, it’s hard to say if you’ve missed defining other hubs. With that many satellites, I’d say there’s at least a decent chance you could find more hubs but it’s not a given.

Yeah I guess that makes more sense about how this would come about naturally. In that case I can see why you’re questioning whether you’ve not got the right hub structure. So it’s not about splitting data it’s about that specificity of terms.

It sounds like your source and your business are speaking on very different terms. Is it that there’s a lot of technical detail in the source system that the business just isn’t interested in? Or at least sees them as attributes rather than full entities in their own right.

Thanks for clarifying the misunderstanding
Frankie

A couple of examples would be helpful for context, without disclosing anything sensitive of course. What is the Hub business key (what is the business entity) and what are a couple of examples of Multi-active Satellites on it?

I work exclusively in healthcare and we run into quite a few examples of data that contains multi-active records. Some of those, as Patrick points out, could be dependent children. In some cases for us it’s a coin toss of which to use based on the data and other factors.