Phillip Cloud comments

Results 993 comments of


                                            Phillip Cloud

bug(postgres): `to_pyarrow` fails with json type

You should be able to use something like ```python pa.Table.from_pandas(expr.to_pandas()) ``` If you definitely need a PyArrow Table.

bug(postgres): `to_pyarrow` fails with json type

Not 100% sure what the ETA is. It will probably be in the next release, but no promises :)

test(duckdb): get things passing with duckdb 0.10.0

I'm going to unwind these cotangent changes and address the inconsistency in a follow-up PR.

test(duckdb): get things passing with duckdb 0.10.0

Well, not a huge fan of merging a dependency that results in a segfault. I don't have a windows machine lying around to test this either.

test(duckdb): get things passing with duckdb 0.10.0

Reported the failure here: https://github.com/duckdb/duckdb/issues/10787

test(duckdb): get things passing with duckdb 0.10.0

I'm going to rehash and force push a couple of times to see if the problem can be reproduced when running tests serially.

test(duckdb): get things passing with duckdb 0.10.0

Looks like something is causing a core dump on Linux now: https://github.com/ibis-project/ibis/actions/runs/7994707013/job/21833372662?pr=8373 😞

test(duckdb): get things passing with duckdb 0.10.0

I guess we can't really test 0.10 in CI until some of these issues are straightened out. I don't really want to be chasing aborts and segfaults.

test(duckdb): get things passing with duckdb 0.10.0

Running tests serially seems to sidestep the windows segfault for now, and the geospatial issue doesn't seem to reproduce anymore. I'll trigger a few more CI runs to see if...

test(duckdb): get things passing with duckdb 0.10.0

@ncclementi @gforsyth @jcrist Anyone opposed to merging this? The windows segfaults don't seem to reproduce under serial testing (not great, but reported upstream) and the linux crashes also do not...