Would it be possible to allow CSV files as a source “database” via the use of a JDBC driver such as http://csvjdbc.sourceforge.net/ ?
You’d need to write a driver for Metabase. Better solution is to just load the csv into a proper database. Faster query times too.
One would not be able to join the csv data with data from other sources.
But one would still be able to create multi dataset curves.
Sorry I do not have any closure skills, so I will be of no help, but again, implementing a Dremio driver would bring valuable features to Metabase : federated sql queries, ‘unmanaged’ data sources (csv, excel, parquet, json), materialized datasets.
Thank you @reydavid for your current work on the subject!
And thank you @dremio for providing us with this great tool.
Best regards from French west coast.
Upvote the following issue:
A solution that currently works is to use Google Sheets. Then you can connect your Sheet to BigQuery which you can then connect to Metabase.
I tested it and it works. Yet I do not like the workaround.
Thank you @riebnik Unfortunately due to data privacy concerns using any cloud based solutions is not an option
Thank you @flamber and @markenson I gave it a try on Ver35.1 got
No method in multimethod ‘connection-details->spec’ for dispatch value: :csv
didnt try to rebuild.
It hasn’t been updated, so it doesn’t work on 0.34+:
The driver has been updated to work with 0.35 as of April 2020.
There’s alternative ways - have a look at the comments:
https://github.com/metabase/metabase/issues/6094 - upvote by clicking on the first post