Phillip Cloud
Phillip Cloud
Clouds are good: ``` …/ibis on bigquery-table-unnest is 📦 v9.1.0 via 🐍 v3.10.14 via ❄️ impure (ibis-3.10.14-env) took 30s ❯ pytest -m 'bigquery or snowflake' -n 8 --dist loadgroup...
This seems to be either a misuse of it by us, or a bug in DuckDB. Here's a test case that reproduces the behavior without Ibis: ```python def test_batch_insert_duckdb(): import...
Ok, interesting. If I create a cursor instead, then I am able to read batches from and insert into a table using the same connection. Poking around to see if...
Ok, so there are a couple issues. Calling `insert` while producing record batches can be addressed by creating a cursor when producing the batches, with the caveat that the new...
I'll put up a PR for discussion.
Ok, I've opened an issue on the DuckDB tracker instead: https://github.com/duckdb/duckdb/issues/12713. Using a cursor inside our `to_pyarrow_batches()` implementation breaks some existing tests and use cases, namely streaming record batches from...
If we move to `cases` then this shouldn't be an issue anymore.
Hey Nick! This is a super interesting problem, but I think it's not feasible for us to correlate an ibis expression with a particular duration (or some proxy of that...
@ExpectationMax Thanks for the issue! Great job on analyzing the problem. Care to submit a PR and we can continue the discussion there?