对 Pandas 列中的一串数字进行排序

Question

I've created previously a python script that creates an author index.
To spare you the details, (since extracting text from a pdf was pretty hard) I created a minimal reproducible example. My current status is I get a new line for each author and a comma separated list of the pages on which the author appears. However I would like to sort the list of pages in ascending manner.

import pandas as pd
import csv
words = ["Autor1","Max Mustermann","Max Mustermann","Autor1","Bertha Musterfrau","Author2"]
pages = [15,13,5,1,17,20]
str_pages = list(map(str, pages))
df = pd.DataFrame({"Autor":words,"Pages":str_pages})
df = df.drop_duplicates().sort_values(by="Autor").reset_index(drop=True)
df = df.groupby("Autor")['Pages'].apply(lambda x: ','.join(x)).reset_index()
df

This produces the desired output (except the sorting of the pages).

               Autor Pages
0            Author2    20
1             Autor1  15,1
2  Bertha Musterfrau    17
3     Max Mustermann  13,5

I tried to vectorize the Pages column to string, split by the comma and applied a lambda function that is supposed to sort the resulting list.

df["Pages"] = df["Pages"].str.split(",").apply(lambda x: sorted(x))
df

However this only worked for Autor1 but not for Max Mustermann . I cant seem to figure out why this is the case

               Autor    Pages
0            Author2     [20]
1             Autor1  [1, 15]
2  Bertha Musterfrau     [17]
3     Max Mustermann  [13, 5]

Answer 1

str.split returns lists of strings. So lambda x: sorted(x) still sort by strings, not integers.

You can try:

df['Pages'] = (df.Pages.str.split(',')
   .explode().astype(int)
   .sort_values()
   .groupby(level=0).agg(list)
)

Output:

               Autor    Pages
0            Author2     [20]
1             Autor1  [1, 15]
2  Bertha Musterfrau     [17]
3     Max Mustermann  [5, 13]

Answer 2

If you want to use your existing approach,

df.Pages = (
    df.Pages.str.split(",")
        .apply(lambda x: sorted(x, key=lambda x: int(x)))
)

---

               Autor    Pages
0            Author2     [20]
1             Autor1  [1, 15]
2  Bertha Musterfrau     [17]
3     Max Mustermann  [5, 13]