Duplicate of #2553

Thanks for pointing this out @mwaskom , sorry about that.

However, the use of common_bins=False (suggested in #2472) does not seem to do the job here; sharex=False now seems to work, however the bin size is now different also across the hue (within the same graph).

And my situation is actually more similar to the following one, where I also have different rows:

import pandas as pd

import seaborn as sns
from matplotlib import pyplot as plt

titanic = sns.load_dataset("titanic")
df = pd.melt(titanic, id_vars=["sex", "class"], value_vars=["fare", "age"])
sns.displot(
    data=df,
    x="value",
    col="variable",
    row="class",
    hue="sex",
    facet_kws={"sharex": False, "sharey": False},
    common_bins=False,
)
plt.savefig("test.png")

Is there a way to share bins across hue and row, but not across col?

Well you can set the bin width you want to use, but for a deeper level of facet-specific parameterization you're going to need to use FacetGrid directly and loop over the facets while passing the specific kwargs you want to use for each one.

Ok, I'll try that. Thanks!

@mwaskom I've been using FacetGrid directly as you suggested above with some success, however I was not able to get a legend based on the hue. Essentially the same outcome described in #2472. Is there no way to get histograms that share axis across the different columns (i.e. same quantities plotted on a row, for different parameters) and have the same bin width, while also getting the legend according to the hue? I haven't been able to achieve this in any way so far. It's not possible at all because of the issues described in #2231?

Hi @RMeli hard to advise without seeing code for your current solution. Could you make a reproducible example and ask on StackOverflow? It's an interesting task. Thanks.

I was about to post some code yesterday but realized it was quite similar to the code above, I think I'm running in circles... I'll post on StackOverflow, thanks.

For future reference: share histogram bins across columns - StackOverflow