Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

While native materialized view feature is a great start, unfortunately they're not useful in a practical way if you have data lineage. They works like a black-box and they can't guarantee the query performance.

The new generation ELT tools such as dbt partially solve this problem. You can model your data and incrementally update the tables that can be used in your BI tools. Looker's Aggregate Awareness is also a great start but unfortunately it only works for Looker.

We try to solve this problem with metriql as well: https://metriql.com/introduction/aggregates The idea is to define these measures & dimensions once and use it everyone; your BI tools, data science tools, etc.

Disclaimer: I'm the tech lead of the project.



Is that related to lightdash.com somehow?

It seems like a very similar technology and also the webpages for both are almost identical.


We both use dbt as the data modeling layer but we don't actually develop a standalone BI tool. Instead, we integrate to third-party BI tools such as Data Studio, Tableau, Metabase, etc.

We love Looker and wanted bring the LookML experience to existing BI tools rather than introducing a new BI tool, that's how metriql was born. I believe that Lightdash is a cool project especially for data analysts who are extensively using dbt but metriql targets users who are already using a BI tool. I'm not particularly sure which pages are identical, can you please point me?


Compare https://metriql.com/introduction/creating-datasets and https://docs.lightdash.com/guides/how-to-create-metrics

I though you are affiliated somehow, but looking at it now, it seems you just use the same documentation website generator :)


Ah, that makes sense! We use https://docusaurus.io/ primarily because dbt uses it as well. :)

It would be great if we can team up to build an open specification for the metric definitions though.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: