将 Pandas Dataframe 中的行与如何处理每一列的特定规则合并

Question

I have a dataframe with meaningless index that contains rows in which one of the columns can be repeated. Something like this:

df = pd.DataFrame({
    'file': ['file1.txt', 'file2.txt', 'file3.txt', 'file1.txt'],
    'size': ['52', '41', '32', '55'],
    'attempts': [4, 4, 3, 1]
})

        file size  attempts
0  file1.txt   52         4
1  file2.txt   41         4
2  file3.txt   32         3
3  file1.txt   55         1

I want to get rid of duplicates, but not just deleting them. I'd like to keep just one row per different 'file', that the 'size' column becomes the maximum all the repeated elements, and that the 'attempts' column becomes the sum of repeated 'attempts'. In other words, I'd like to get:

        file size  attempts
0  file1.txt   55         5
1  file2.txt   41         4
2  file3.txt   32         3

I know how to do this explicitly looping through the Dataframe, but I'd like to make it more efficient.

Answer 1

You can use .groupby() and .agg() to aggregate with size using max value and attempts using sum on the attempt counts.

df.groupby('file', as_index=False).agg({'size': 'max', 'attempts': 'sum'})

Result:

        file size  attempts
0  file1.txt   55         5
1  file2.txt   41         4
2  file3.txt   32         3

Answer 2

This is a very dirty solution, but nonetheless should work:

df_ = df.groupby('file').max('size')
df_ = pd.concat([df_.iloc[:,0], df.groupby('file').sum('attempts')['attempts']], axis=1)

df_

           size     attempts
file        
file1.txt   55         5
file2.txt   41         4
file3.txt   32         3