BUG: df.style.apply(subset=...) argument only works for explicit slices but not ":"

[dieterwang]

Related to #25858 (if not the same?), reproduced on 0.25.3 and current version 1.0.3.

It seems that the subset argument for pd.DataFrame.style only works for slices with explicit labels, e.g. ['A',2] but not for select-all labels, e.g. [:,2]. See MWE below

import pandas as pd
pd.__version__ # 1.0.3

# Taken from 
# https://pandas.pydata.org/pandas-docs/stable/user_guide/style.html
def color_negative_red(val):
    color = 'red' if val < 0 else 'black'
    return 'color: %s' % color

df = pd.DataFrame([
    [-1, 3, 6],
    [-2,-3, 2]
])
df.index = pd.MultiIndex.from_tuples([('A',1),('A',2)])

# Style works for applymap
df.style.applymap(color_negative_red)

# Works for explicit slice
slicer = pd.IndexSlice[pd.IndexSlice['A',2],:]
df.loc[slicer] # valid .loc slicer
df.style.applymap(color_negative_red, subset=slicer) # works

# Breaks for select-all slice
slicer = pd.IndexSlice[pd.IndexSlice[:,2],:]
df.loc[slicer] # valid .loc slicer
df.style.applymap(color_negative_red, subset=slicer)
# TypeError: unhashable type: 'slice'

[attack68]

This no longer seems to error on master: 1.2.0.dev0+186.g25a895cf0

[jreback]

@attack68 happy to take a test for this

[jimprince]

Also getting this error but when I'm using the background_gradient method of style.

I started off on 'pandas 1.1.14' -> Issue found, log attached
Upgraded to '1.2.0.dev0+1072.g31db14c48' -> issue persists, log attached
Downgraded to 'v1.2.0.dev0-186-gf46ebe895' to try out approximately when @attack68 says this was working. -> Issue persists.

My code is:

import pandas as pd
import numpy as np

def mklbl(prefix, n):
    return ["%s%s" % (prefix, i) for i in range(n)]

miindex = pd.MultiIndex.from_product([mklbl('A', 5),['D', 2, "x", "J"]])
micolumns = pd.Index(['a', 'foo', 'b', 'bar'])
dfmi = pd.DataFrame(np.arange(len(miindex) * len(micolumns))\
                    .reshape((len(miindex), len(micolumns)))\
                    ,index=miindex,\
                    columns=micolumns\
                   ).sort_index()\
                   .sort_index(axis=1)

slicer = pd.IndexSlice[pd.IndexSlice[:, "D"], :]
print(dfmi.loc[slicer])
dfmi.style.background_gradient(cmap="Red", subset=slicer)

I have a jupyterlab docker container pulled from the official repo running on my MBP.

Thanks in advance to everyone!

Pandas_output1.1.4.txt
Pandas_output1.2.0.dev01072.g31db14c48.txt