7 Comments

For me, in a very simple way, the metrics layer is a subset of a semantic layer - as you pointed out where the semantic layer is covering a lot more use cases (like entities).

Expand full comment
author

Yes it’s more like a single access point for data

Expand full comment

What is in practice the difference between a semantic/metric layer and the old datamarts?

Expand full comment
author

The data marts are materialisations of your data model. The semantic layer provides an interface to use your data model without needing knowledge of it.

Expand full comment

Thanks for replying back David. But what exactly does that means in practice?

- Is a semantic layer a combination of tables or views that are exposed via API? Is it something else?

- Are DataMarts part of the Semantic Layer?

- Is the Semantic Layer coming after the dimensions and measures from the DataWarehouse?

I mean, what exactly is the semantic layer?

Thank you

Expand full comment
author

The semantic layer allows you to access data from that data model materialised as tables.

It defines what the entities in the data model are in real world terms and how to access them.

Yes they often come with APIs to easily know what the semantic layer contains and request this data, but that is kind of a necessary bonus.

Check out cube.dev it’s a very good example

Expand full comment

Thanks a lot David. I will.

Expand full comment