Thijs

Not exactly the same, Nessie is a bit special in that it has the ability to have NONE as authentication type https://projectnessie.org/nessie-latest/client_config/#authentication-type-none

I believe the architecture for this is implemented now with #230 Only the partition evolution remains, as well as supporting the remaining transforms

I'm closing this as stale, we've had multiple releases since this issue was last commented on. Feel free to reopen if the issue persists

> i have a similar example I don't think these are the same problem. @lamb-russell your issue likely predates [this](https://github.com/duckdb/duckdb/pull/15563) PR, adding support for reading paths that start with `file://`...

So I don't forget, this needs a follow-up PR to add `duckdb` -> `substrait` -> `ibis` verification. This PR is for `ibis` -> `substrait` -> `duckdb`

Seems something happened to timestamp extract support since I made this PR? ```py duck = duckdb.connect(db_path) duck.execute(""" create table tbl(date timestamp) """) duck.execute(""" insert into tbl values ('2021/09/21 12:02:21'::TIMESTAMP) """)...

+1 for random, when there are many results for a search it can be useful to randomize them

@samansmink I assume our integrated HFFS doesn't allow writes yet? A workaround would be to attach the hffs filesystem as you would before It should take priority over our own...

Can you post what `make` command you're running to build python?

As a follow up, we can rework the vector size and block size tests we have that rely on manually editing the files I've committed those changes by accident more...