How can I chain the conditions for style functions in pandas?

Question

That's what I want to achieve

I do have a Pandas Dataframe - e.g. this one:

            0           1           2
0  110.718803  119.821042  -52.593518
1   65.180254   33.518722  -69.893688
2 -135.652788  -64.711718 -241.717819
3  237.781393  -56.865142   15.969767
4  141.585158  138.904568 -115.155063
5   10.030938  -59.274415   73.328127
6  106.937681   -3.604859   44.418938
7  -49.478211  -91.574908  160.340627
8  170.744019  -85.764809  246.141857
9  -94.246832   81.069700 -113.460438

Based on 3 conditions the background-color of the cell should be different:
cell <= 0 should be in red
cell >= 100 should be in blue
all other cells

That's what I did to achieve that

I wrote this function (based on the infos in Pandas documentation Pandas styling:

def highlight_number(row):
    return [
        'background-color: red; color: white' if cell <= 0
        else 'background-color: green; color: white'
        for cell in row
    ]


df.style.apply(highlight_number)

it works fine for two conditions.

That's my problem

I tried different methods to add the third condition into the function above but I always got back an error.
Could you please give me a hint how to add the condition inside the list?
I didn't find an answer. Thx a lot.

Answer 1

You can write a normal for loop instead of list comprehension.

def highlight_number(row):
    arr = []
    for cell in row:
        if  cell <= 0:
            arr.append('background-color: red; color: white')
        elif cell >= 100:
            arr.append('background-color: blue; color: white')
        else:
            arr.append('background-color: white; color: black')
    return arr
df.style.apply(highlight_number)

Output:

Answer 2

easiest way I feel is using np.select which takes in a list of conditions and a list of choices (like if elif and also takes a default value like else) and modifying your function and use it over axis=None since we are applying to the entire dataframe (note you can also use subset for a subset of columns)

def highlight_number(dataframe):
    
    conditions = [dataframe <=0,(dataframe>0) & (dataframe>=100)] #your conditions
    choices = ['background-color: red; color: white',
               'background-color: blue; color: white']
    arr = np.select(conditions,choices,'background-color: white; color: black')
    return pd.DataFrame(arr,index=dataframe.index,columns=dataframe.columns)


df.style.apply(highlight_number,axis=None)

---