Geoff Genz

What `clickhouse_connect` version are you using? This was fixed in v0.6.11 and I just ran tests with Pandas 2.2.1 in my local environment.

@Kevin9436 That's a great datapoint. To achieve the same results with Native format in clickhouse-connect will be very challenging, since the parquet approach offloads data transformation (from ClickHouse format to...

Thanks for the schema. I'm fairly sure we could do some Pandas optimizations with either Arrow or low level numpy to improve the query_df for that kind of basic datatypes,...

Agreed, this should have had its own enhancement request/issue before now.

The `await/async` keywords are part of the Python standard `asyncio` library and work fine with `clickhouse-connect`. The main caveat is that ClickHouse doesn't allow concurrent queries in the same session,...

To clarify, I see this issue as an enhancement request to either create a clean async wrapper around the library that feels more "async native" or even better, use a...

@alexted We in the process of defining the roadmap for `clickhouse-connect` and other integrations, but for the moment this work is not planned for the immediate future. The current thinking...

I've added an [example](https://github.com/ClickHouse/clickhouse-connect/blob/main/examples/run_async.py) of how to run `clickhouse-connect` queries asynchronously (it includes a semaphore just for fun as well). Based on very limited testing, this solution still significantly outperforms...

@samuelcolvin Putting aside the Arrow Flight question for the moment, I'm not sure what problem you are trying to "solve". Queries run using the wrapper yield the asyncio event loop...

We don't currently test or support Pandas 2.0, but in fact the new datatypes are technically more correct.