There are some limitations with the implementation currently due to the lack of INFORMATION_SCHEMA and being restriction to JSON types only, but I'm glad you think it can be useful 🙂 we are definitely keen to figure out whether it can be pushed forward and rolled out more officially at some point - thanks for the feedback!

Missed those restrictions. So every "column" looks like JSON?

And by information_schema you mean that tools cannot see for example what tables exist, and thus you cannot use any explorative tools? That's actually ok for now, we can manage with specific queries

yes and yes

actually a lot of Postgres-friendly BI tools will look for pg_catalog (not INFORMATION_SCHEMA per the SQL standard), which would likely lead us down an additional rabbit hole, like CockroachDB fell into https://www.cockroachlabs.com/docs/stable/pg-catalog all the tradeoffs going down this route will be very nuanced

ideally FlightSQL can close the gap on functionality, and offer superior performance all round

Does anything support FlightSQL yet?

Heh, looks like postgres does

oh cool! I missed that 🙂 also: https://docs.dremio.com/current/sonar/developing-client-apps/arrow-flight-sql/ https://github.com/voltrondata/flight-sql-server-example https://www.postgresql.org/about/news/apache-arrow-flight-sql-adapter-for-postgresql-010-2716/ https://medium.com/learning-the-computers/test-driving-the-snowflake-adbc-driver-6671c373b008

Referring to your columnar data webinar, does flightsql handle it all, or is it best left for OLAP and still use something else (pgwire etc) for transactional queries?

FlightSQL can potentially evolve to handle it all, it's built on grpc so should be very good foundation for small & fast transactions/queries too, but ~all the vendor activity looks to be OLAP focused so far

Ok, interesting. If only bitemporality would also get everywhere so that you could really have all things you need in one package

Also, I'm still stuck in the mindset that Apache means the httpd. And now I can't even keep track of all their projects only around databases/analysis

> There are currently 300+ open source initiatives at the ASF Oh wow

66 in big data category

Didn't it used to be that databases were ancient and you only really had a couple of choices, even less if you don't have money, and every week there's a JS framework. Now we've had React forever and a new data product every week

hold on https://react.dev/reference/react are you talking about 🙂

Whatever all the clojure libs use 😉

Here's the list of things being worked on currently: https://docs.xtdb.com/intro/roadmap.html