汇总后排序并选择（熊猫）

Question

I would like to aggregate a Pandas DataFrame in order to count the number of children (variable child_name ) for each father (variable father_name ). The dataframe looks like this (it is a toy example of course, I want to grasp the concept):

father_name   child_name
Robert        Julian
Robert        Emily
Robert        Dan
Carl          Jack
Carl          Rose
John          Lucy
Paul          Christopher
Paul          Thomas

Now, I define an aggregation dictionary and use it on the dataframe d :

import pandas as pd
aggregation = {
    'child_name': {
        'n_children': 'count'
    }
}
d.groupby('father_name').agg(aggregation)

I obtain this output:

            child_name
            n_children
father_name           
Carl                 2
John                 1
Paul                 2
Robert               3

and now I would like to:

• sort the fathers according to their number of children (in decreasing order)
• show only the fathers that have 2 or more children

How can I do that? Maybe there's also a quicker way to do this, but I would like to learn this method too. Thanks in advance!

Answer 1

You could let

df_count = df.groupby('father_name').count()
df_count[df_count.child_name > 1].sort_values(by='child_name', ascending=False)

Output:

             child_name
father_name
Robert                3
Carl                  2
Paul                  2

If you want to make heavier use of agg , that might look something like the following (which will throw a FutureWarning as renaming using dicts is deprecated):

df.groupby('father_name').agg({'child_name': {'n_children': lambda x: len(x) if len(x) > 1 else None}}).dropna()

then sorting the result afterwards.

Answer 2

Let's try like this way to meet your two conditions-

    import pandas as pd
    df = pd.DataFrame({"father_name":["Robert","Robert","Robert","Carl","Carl","John","Paul","Paul"],"child_name":["Julian","Emily","Dan","Jack","Rose","Lucy","Christopher","Thomas"]})

    #sort the fathers according to their number of children (in decreasing order)
    df = df.groupby(by='father_name').count().sort_values(['child_name'],ascending=False)

    #show only the fathers that have 2 or more children
    df_greater_2 = df[df['child_name'] >= 2]

    print(df_greater_2)

DEMO: https://repl.it/@SanyAhmed/EarnestTatteredRepo