I’ve been looking for a while for an open-source data dictionary/catalog/glossary. Everything I’ve found has been pretty disappointing. I know that metabase isn’t explicitly intended to be a data dictionary, however it seems that with the ability it has to automatically ingest a schema, and then let users add descriptive information to the data, it’s got most of the basics already.
One core feature that would make it work is the ability to search the metadata. I know you can search the question library, but what I really am after is the ability to search for tables and columns, and their descriptions, for keywords. Eg, find me all the tables that contain the field ‘user_id’, or where the description mentions ‘sales’ etc.
Any plans to add such a feature? I couldn’t see anything in the github issues. I’m interested in contributing if anyone else agrees it’s a good addition.
It seems @sbelakjust started working on a backend feature that could also help with discoverability and exploration of a datamodel - so IMO complement a metadata search feature.
I’m also extremely interested in this topic. I guess you still might#3089 be? (if the two @EvilPuppetMaster entities are the same )
Some related efforts I have come across:
SchemaSpy - note among other features it can extract database schema comments.
Databook - still very young. Effort was introduced on the Apache Airflow mailing list (sadly I’m currently just another lurker there)
I would like such a feature as well. at the minimum - does anyone have the sql queries i can run on the metabase data repository to look at all the descriptions, comments people have added to every table, field, metric, question?
Yep I’m still interested. I’ve contributed a PR (https://github.com/metabase/metabase/pull/6974) that allows comments from tables and columns to be synced as the initial descriptions in Metabase. I’m thinking of tackling the metadata search next. I’ll check out the issue and other projects you mentioned @jornh, thanks!
)