基于groupby中的多个列值创建分类列

Question

print(df.groupby(['Step1', 'Step2', 'Step3']).size().reset_index(name='Freq'))

                       Step1                    Step2                   Step3         Freq
0                       6.0                     17.6                    28.60         135
1                       7.5                     22.0                    35.75         255
2                      10.5                     30.8                    50.05         129
3                      12.0                     35.2                    57.20         369
4                      13.5                     39.6                    64.35         249
5                      15.0                     44.0                    71.50         246
6                      16.5                     48.4                    78.65         246
7                      18.0                     52.8                    85.80         369
8                      21.0                     61.6                   100.10         375
9                      22.5                     66.0                   107.25         249
10                     25.5                     74.8                   121.55         123

The 'Step1', 'Step2', 'Step3' columns are constant input values. There are 10 unique combinations of input values from these columns (shown in the groupby). I am looking to delete the individual 'Step1', 'Step2', 'Step3' columns and create a single column "Step Type" that has a letter that represents the unique combinations of input values from these columns.

Desired output:

                     Step Type   Freq
0                      A         135
1                      B         255
2                      C         129
3                      D         369
4                      E         249
5                      F         246
6                      G         246
7                      H         369
8                      J         375
9                      L         249
10                     M         123

Step Type A: Step1=6.0, Step2=17.6, Step3=28.60

How do I do this?

Answer 1

As combinationss of the three steps are unique, I used the combinations as a key of Dictionary for Step Type.

Here, I pre-defined category value but it can be auto-generated by scanning the df if needed.

# df
   Step1  Step2  Step3
0    6.0   17.6  28.60
1    7.5   22.0  35.75
2   10.5   30.8  50.05
3   12.0   35.2  57.20
4   13.5   30.6  64.35

category = {
    (6.0, 17.6, 28.60): 'A',
    (7.5, 22.0, 35.75): 'B',
    (10.5, 30.8, 50.05): 'C',
    (12, 35.2, 57.20): 'D',
    (13.5, 30.6, 64.35): 'E',
}

df['Step_Type'] = df.apply(lambda row: category[(row['Step1'], row['Step2'], row['Step3'])], axis=1)

df = df[['Step_Type', 'Freq']]
print(df)

#  Step_Type  Freq
#0         A   135
#1         B   255
#2         C   129
#3         D   369
#4         E   249