Paul Hobson

Thanks for the report. Any interest in submitting a PR to fix the documentation?

Thanks for the precisely described and demonstrated issue. I think this stems from a gap in the documentation. A PP plot doesn't actually use the distribution of the data. It's...

@ev-br I love that even someone has interest in paramnormal. I can't promise that I'll be able to get to everything very quickly -- but I'm sure you understand how...

Excellent list.

@jrbourbeau I left the xfail mark on the test parameter, so it ["successfully failed"](https://github.com/coiled/dask-snowflake/actions/runs/3596014405/jobs/6056194174#step:5:66) here. Sorry for that confusion. I'm trying to (roughly) bisect where things fall apart. I'll push...

@jrbourbeau (cc @hayesgb) I figured out what's going on here. Since we have so little control over the sizes of the batches that snowflake is returning, so of the individual...

Potential third option from me: relax the test a bit. Current the check is: ```python assert (partition_sizes < 2 * parse_bytes("2 MiB")).all() ``` So we're already using a fudge factor...

@inkenbrandt this has been on my TODO list for a while, but I haven't gotten there yet. It's a sad state of affairs when the tests are the best examples...

@inkenbrandt I'm thinking I should remove the ability to return dataframes. I think a best if give the wrong impression about the nature of the imputation. Thoughts?

Thanks for the clear demonstration of the issue. Is "phone" your parquet dataset in this case, or is it a directory of different parquet datasets that happen to have the...