lmeyerov

@vaimdev can you update from `master` & retest , esp b/c changes from https://github.com/graphistry/pygraphistry/pull/615 @DataBoyTX any more to do to land? this has been drifting and seems worth putting in?

https://github.com/xzkostyan/clickhouse-sqlalchemy/blob/3dc8df9da598ac51e20e9b7bb110ae97e250fab7/setup.py#L99 maybe we just need to rerun the conda-forge feedstock? I'm unsure how that works: https://github.com/conda-forge/clickhouse-sqlalchemy-feedstock

I think the validators can catch this, but probably a deeper backend issue we should explore for cases where validators aren't run, cc @aucahuasi

Bringing this back up as useful for gfql , loiue, etc, as came up in a customer call today (`gfql` user) -- * In PyGraphistry side, probably something like: *...

Implementation note: choose predicates that stay vectorized for pandas/cuDF (avoid Python loops). For list/set columns, aim for built-ins like cudf // equivalents and pandas avoidance—ideally via +groupby or -style ops—so...

Likely other calls that trigger upload, like upload() Unclear how deep / where to thread this Also, other ways graphs can be dirty ... Should this scope to just pd-->...

Ah that makes sense -- so `"x": "b"` is sugar that we turn into `"x": eq("a")`, and add more type polymorphism to `eq()` Just double checking: did you encounter this...

Hi @morristai thanks for the pictures and example to help discuss 1. I'm surprised the results are different tbh, and suspect a bug in the first case, as Tina should...

I confused myself in the reply -- current n-hop is already `1..n`: so inclusive, not 'exactly n hops'. That is why tina (n=2) was allowed in case 1, and rejected...

I'm not seeing the path to Tina that you'd expect to be in the match set -- I only see a 2 hop and 4+ hop paths, while the query...