Categorical order in hierarchical axis not respected

[flying-sheep]

ALL software version info

(this library, plus any other relevant software, e.g. bokeh, python, notebook, OS, browser, etc should be added within the dropdown below.)

Software Version Info

Using Python 3.12.7 environment at /home/phil/.local/share/hatch/env/virtual/dask millions of cells/cMqPOBoB/dask millions of cells
Package            Version
------------------ -----------
asttokens          2.4.1
bleach             6.2.0
bokeh              3.6.1
certifi            2024.8.30
charset-normalizer 3.4.0
colorcet           3.1.0
comm               0.2.2
contourpy          1.3.1
cycler             0.12.1
debugpy            1.8.8
decorator          5.1.1
executing          2.1.0
fonttools          4.55.0
holoviews          1.20.0
idna               3.10
ipykernel          6.29.5
ipython            8.29.0
ipywidgets         8.1.5
jedi               0.19.2
jinja2             3.1.4
jupyter-client     8.6.3
jupyter-core       5.7.2
jupyterlab-widgets 3.0.13
kiwisolver         1.4.7
linkify-it-py      2.0.3
markdown           3.7
markdown-it-py     3.0.0
markupsafe         3.0.2
matplotlib         3.9.2
matplotlib-inline  0.1.7
mdit-py-plugins    0.4.2
mdurl              0.1.2
mizani             0.13.0
nest-asyncio       1.6.0
numpy              2.1.3
packaging          24.2
pandas             2.2.3
panel              1.5.4
param              2.1.1
parso              0.8.4
patsy              1.0.1
pexpect            4.9.0
pillow             11.0.0
platformdirs       4.3.6
plotnine           0.14.1
prompt-toolkit     3.0.48
psutil             6.1.0
ptyprocess         0.7.0
pure-eval          0.2.3
pygments           2.18.0
pyparsing          3.2.0
python-dateutil    2.9.0.post0
pytz               2024.2
pyviz-comms        3.0.3
pyyaml             6.0.2
pyzmq              26.2.0
requests           2.32.3
scipy              1.14.1
six                1.16.0
stack-data         0.6.3
statsmodels        0.14.4
tornado            6.4.1
tqdm               4.67.0
traitlets          5.14.3
typing-extensions  4.12.2
tzdata             2024.2
uc-micro-py        1.0.3
urllib3            2.2.3
wcwidth            0.2.13
webencodings       0.5.1
widgetsnbextension 4.0.13
xyzservices        2024.9.0

Description of expected behavior and the observed behavior

When using multiple kvars, the category order is ignored:

Complete, minimal, self-contained example code that reproduces the issue

import pandas as pd
import holoviews as hv
hv.extension('bokeh')

cells_dtype = pd.CategoricalDtype(pd.array(["~1M", "~10M", "~100M"], dtype="string"), ordered=True)

df = pd.DataFrame(dict(
    cells=cells_dtype.categories.astype(cells_dtype),
    time=pd.array([2.99, 18.5, 835.2]),
    function=pd.array(["read", "read", "read"]),
))

hv.Bars(df, ["function", "cells"], ["time"])

Stack traceback and/or browser JavaScript console output

>>> df["cells"]
0      ~1M
1     ~10M
2    ~100M
Name: cells, dtype: category
Categories (3, string): [~1M < ~10M < ~100M]

Screenshots or screencasts of the bug in action

• I may be interested in making a pull request to address this

[hoxbro]

Sounds reasonable. I think the correct place to implement this is around here:

holoviews/holoviews/element/util.py

Lines 147 to 164 in c227510

	for group in grouped:
	vals = group.dimension_values(ydim, False)
	if len(vals) == 1:
	orderings[vals[0]] = [vals[0]]
	else:
	for i in range(len(vals)-1):
	p1, p2 = vals[i:i+2]
	orderings[p1] = [p2]
	if sort:
	if vals.dtype.kind in ('i', 'f'):
	sort = (np.diff(vals)>=0).all()
	else:
	sort = np.array_equal(np.sort(vals), vals)
	if sort or one_to_one(orderings, ycoords):
	ycoords = np.sort(ycoords)
	elif not is_cyclic(orderings):
	coords = list(itertools.chain(*sort_topologically(orderings)))
	ycoords = coords if len(coords) == len(ycoords) else np.sort(ycoords)