Computed number field for data model relations

[Taldres]

Scope & Context

It would be really cool to have a kind of calculated sum in the data models, which aggregates the values of a specific field from all entries of a model that are related to this one.

Like the number of machines from multiple production plants that belong to the model.

Current behavior

I added a Machines number field to the Company model, which defines the total number of machines within a company. To get more detailed and structured information, I created a Production Site model, which belongs to a Company and also has a Machines number field.

Currently, I have to set the total number of machines manually.

Expected behavior

A computed number or sum number field that I could link to my related model to calculate the expected total amount.

[Sandheep-OSC]

This looks interesting to me. I would like to work in it.

[charlesBochet]

Hi @Taldres, this is indeed something we have in mind.

The product still needs to be refined about it.
This is touching two areas we are still refining: Computed Fields and aggregations.

We will introduce a new type of field we would call a "Computed Field" (or a "LookUp Field"). The tricky part is that we cannot compute this at read without introducing major performance issues on index views ; this would indeed force the engine to join additional tables to get the data and can slow down the app very quickly. Instead we would probably want to compute these Computed Field at write time and would probably leverage the workflow/automation work that is ongoing

Aggregations: regarding your specific example, it also reminds me of aggregations. We will likely introduce new rest / graphql endpoints to get basic aggregates on tables (sum, average, ...) and add the possibility to query them while querying relations. Here we could for example allow to display the number of "machines" next to the production sites

I think it's still a bit early for Twenty to tackle this but it's definitely something we have in mind

[Taldres]

Hi @Taldres, this is indeed something we have in mind.

The product still needs to be refined about it. This is touching two areas we are still refining: Computed Fields and aggregations.

We will introduce a new type of field we would call a "Computed Field" (or a "LookUp Field"). The tricky part is that we cannot compute this at read without introducing major performance issues on index views ; this would indeed force the engine to join additional tables to get the data and can slow down the app very quickly. Instead we would probably want to compute these Computed Field at write time and would probably leverage the workflow/automation work that is ongoing

Aggregations: regarding your specific example, it also reminds me of aggregations. We will likely introduce new rest / graphql endpoints to get basic aggregates on tables (sum, average, ...) and add the possibility to query them while querying relations. Here we could for example allow to display the number of "machines" next to the production sites

I think it's still a bit early for Twenty to tackle this but it's definitely something we have in mind

Thank you for your detailed response. I'm glad to hear that you already have this in mind. I'm looking forward to the future solution.

[FelixMalfait]

Thanks for the suggestion! Converting this to a discussion as this is planned on the roadmap. I hope we'll have something powerful by end of Q1 next year.