pandas 如何拆分给定字段两次

Question

我想统计专栏中排名靠前的主题。有些字段有逗号或点我想用它们创建一个新行。

import pandas as pd
from pandas import DataFrame, Series

sbj = DataFrame(["Africa, Business", "Oceania", 
    "Business.Biology.Pharmacology.Therapeutics", 
    "French Litterature, Philosophy, Arts", "Biology,Business", ""
     ])
sbj

我想拆分成一个新的带有“.”的任意字段或“.”

sbj_top = sbj[0].apply(lambda x: pd.value_counts(x.split(",")) if not pd.isnull(x) else pd.value_counts('---'.split(","))).sum(axis = 0)
sbj_top

我在尝试重新拆分 ('.') 时遇到错误 (AttributeError)

sbj_top = sbj_top.apply(lambda x: pd.value_counts(x.split(".")) if not pd.isnull(x) else pd.value_counts('---'.split(","))).sum(axis = 0)
sbj_top

我想要的输出

sbj_top.sort(ascending=False)
plt.title("Distribution of the top 10 subjects")
plt.ylabel("Frequency")
sbj_top.head(10).plot(kind='bar', color="#348ABD")

Answer 1

您可以将 Counter 与 itertools 中的 chain 一起使用。请注意，我在解析之前先将句点替换为逗号。

from collections import Counter
import itertools
from string import whitespace

trimmed_list = [i.replace('.', ',').split(',') for i in sbj[0].tolist() if i != ""]
item_list = [item.strip(whitespace) for item in itertools.chain(*trimmed_list)]
item_count = Counter(item_list)

>>> item_count.most_common()
[('Business', 3),
 ('Biology', 2),
 ('Oceania', 1),
 ('Pharmacology', 1),
 ('Philosophy', 1),
 ('Africa', 1),
 ('French Litterature', 1),
 ('Therapeutics', 1),
 ('Arts', 1)]

如果您需要以 DataFrame 的形式输出：

df = pd.DataFrame(item_list, columns=['subject'])
>>> df
           subject
0               Africa
1             Business
2              Oceania
3             Business
4              Biology
5         Pharmacology
6         Therapeutics
7   French Litterature
8           Philosophy
9                 Arts
10             Biology
11            Business

pandas 如何拆分给定字段两次

pandas how to split twice a given field

python-2.7

pandas

ipython-notebook