Development workflow//testing

It would be really helpful if there were a feature where you could point models to a development schema when in development mode (e.g. pointing to dynamic dbt dev schemas using a users parameter), and being able to test all models for validity.

Hi @Alex_H, if I understand you correctly, basically you want to config the environment wherein dev mode, all of your models, and datasets are forced to be pointed to the dev database. Whereas, after deploying the project to the Production, all of your deployed models, and datasets will automatically be pointed to the Production database.
This is something we call environment management and we have thought of supporting it in the future. For now, I don’t have the timeline for it yet but I will +1 priority for this.


For this one, could you share a little bit more details about the use case and workflow that you’re thinking of? Also, how do you currently test your models?

1 Like

That’s good to hear environement management is on the radar.

For this one, could you share a little bit more details about the use case and workflow that you’re thinking of? Also, how do you currently test your models?

Once we have defined our measures and dimensions in Holistics, it is possible that during development, a column is removed or renamed. Once the development table is ready, it would be good to be able to automatically test that all the dimensions and dimensions in a model still work. The way to do this now would be to select all the fields in a model and manually run it.

In addition, it would be really helpful if we could view dashboards using the development data (when the dev switch is turned on in AML).

1 Like

If that’s the case, we will always test the validity of your modeling setup before deploying it to production. If your models or datasets are invalid, the deployment will be failed and you will need to resolve it before continuing deploying.


For this one, the use case is valid. That’s the reason we’re working on dashboard as-code so that you can create and test your dashboard in the development area before deploying it to production. I will let you know when there is an update on this.

If that’s the case, we will always test the validity of your modeling setup before deploying it to production. If your models or datasets are invalid, the deployment will be failed and you will need to resolve it before continuing deploying.

But in this scenario nothing would have changed on the Holistics layer - only in the underlying db…

Ok, I see. How frequently does your underlying table structure change?

Difficult to say in the long run as we’re just setting up so they’re changing fairly frequently. My experience though is that models do continuously evolve, but slowly.