Replies: 2 comments
-
|
My take here is that the purpose of the current semantic model is to map more to the size of a data mart (a number of tables to answer as set of questions). Not to map to the size of all the tables in an org. I imagine there is some ontology layer or introspection of semantic models to combine the across data marts. Things that share dimensions or describe what the safe join patterns are. Another direction to think about is composition. Although, the way I think of that is that it helps with re-use, but I think will run into some problems if you try using it to make models that are too large. |
Beta Was this translation helpful? Give feedback.
-
|
Started a related discussion which can help in organizing the semantic models and the underlying physical tables. The model needs to move beyond a single file, even if we consider multiple semantic_models, sharing peices of information between them is crucial. On the same line, using a templating language (another discussion) would help a lot as the inventory grows and needs governance. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
Semantic model has a set of datasets, metrics.
Examples show a single file.
A model with 100s of datasets would likely store them in different files. Is this how the spec intends it to be defined?
Shouldn't this be in the spec (example, where do we put the relations)?
Beta Was this translation helpful? Give feedback.
All reactions