Hi Folks,
A disclaimer on the previous message: While an SQL querying endpoint has
been added to Druid, the computation-engine is still very specific and by
no mean a general-SQL engine. You'll probably very quickly find the limits
of querying Druid via SQL.
For a full SQL engine use Presto :)
Joseph
On Wed, Apr 8, 2020 at 8:08 AM Luca Toscano <ltoscano(a)wikimedia.org> wrote:
Hi everybody,
I recently learned from Superset Upstream that the suggested way to query
Druid is using the SQLAlchemy connector, not the Druid one (that is
"deprecated"). I digged a little bit more about this and I found the
following:
- Druid Analytics supports SQL querying on the brokers, so it was a matter
of adding a new "database" called "Druid Analytics SQL" to Superset.
- "Druid Analytics SQL" can be used for data querying and exploration in
the SQL Lab (feel free to test it!).
- Using SQLAlchemy should be more quick/convenient since we wouldn't need
to define mapping for Druid datasources manually (need to better
investigate this but I am reasonably sure).
Moving all the charts already in Superset to the new database/datasource
is probably a ton of work, and we are not really in a hurry (the
"deprecated" above means that they still support it but a bug to that code
will be considered not a blocker), but what I'd ask it to start using it
and see if there are pros/cons in your daily workflows.
I created a task to report thoughts/suggestions/bugs/etc.. to avoid
spamming too many people:
https://phabricator.wikimedia.org/T249681
Thanks!
Luca
_______________________________________________
Analytics mailing list
Analytics(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics
--
Joseph Allemandou (joal) (he / him)
Sr Data Engineer
Wikimedia Foundation