Pandas - 出现字符串时拆分数据并创建列

Posted 2023-03-12

【中文标题】Pandas - 出现字符串时拆分数据并创建列

【英文标题】：Pandas -Split data and create columns when string occurs

【发布时间】：2021-07-08 14:24:07

【问题描述】：

我希望读取文本文件（见下文），然后仅为所有英语联赛创建列。因此，我将尝试执行类似“别名”为“England_”的操作，然后创建一个以别名作为标题的新列，然后在行中创建玩家名称。请注意，Alias 的第一次出现在文本文件中以“Aliases”的形式出现。

"-----------------------------------------------------------------------------------------------------------" 
"-                                            NEW TEAM                                                    -" 
"-----------------------------------------------------------------------------------------------------------" 
Europe Players
17/04/2019
07:59 p.m.

Aliases for England_Premier League

-------------------------------------------------------------------------------
Harry Kane
Mohamed Salah
Kevin De Bruyne

The command completed successfully.

Alias name     England_Division 1
Comment        Teams

Members

-------------------------------------------------------------------------------
Will Grigg
Jonson Clarke-Harris
Jerry Yates
Ivan Toney
Troy Parrott
The command completed successfully.

Alias name     Spanish La Liga
Comment        

Members

-------------------------------------------------------------------------------
Lionel Messi
Luis Suarez
Cristiano Ronaldo
Sergio Ramos
The command completed successfully.

Alias name     England_Division 2
Comment        

Members

-------------------------------------------------------------------------------
Eoin Doyle
Matt Watters
James Vughan
The command completed successfully.

这是我目前关于如何读取数据的代码

df = pd.read_csv(r'Desktop\SampleData.txt', sep='\n', header=None)

这给了我一个带有一列的 pandas DF。我对python相当陌生，所以我想知道如何获得以下结果？读取文件时应该使用分隔符吗？

England_Premier League	England_Division 1	England_Division 2
Harry Kane	Will Griggs	Eoin Doyle
Mohamed Salah	Jonson Clarke-Harris	Matt Watters
Kevin De Bruyne	Ivan Toney	James Vughan
	Troy Parrott

【参考方案1】：

您可以使用re 模块来完成任务。例如：

import re
import pandas as pd


txt = """
"-----------------------------------------------------------------------------------------------------------" 
"-                                            NEW TEAM                                                    -" 
"-----------------------------------------------------------------------------------------------------------" 
Europe Players
17/04/2019
07:59 p.m.

Aliases for England_Premier League

-------------------------------------------------------------------------------
Harry Kane
Mohamed Salah
Kevin De Bruyne

The command completed successfully.

Alias name     England_Division 1
Comment        Teams

Members

-------------------------------------------------------------------------------
Will Grigg
Jonson Clarke-Harris
Jerry Yates
Ivan Toney
Troy Parrott
The command completed successfully.

Alias name     Spanish La Liga
Comment        

Members

-------------------------------------------------------------------------------
Lionel Messi
Luis Suarez
Cristiano Ronaldo
Sergio Ramos
The command completed successfully.

Alias name     England_Division 2
Comment        

Members

-------------------------------------------------------------------------------
Eoin Doyle
Matt Watters
James Vughan
The command completed successfully.
"""

r_competitions = re.compile(r"^Alias(?:(?:es for)| name)\s*(.*?)$", flags=re.M)
r_names = re.compile(r"^-+$\s*(.*?)\s*The command", flags=re.M | re.S)

dfs = []
for comp, names in zip(r_competitions.findall(txt), r_names.findall(txt)):
    if not "England" in comp:
        continue
    data = []
    for n in names.split("\n"):
        data.append(comp: n)

    dfs.append(pd.DataFrame(data))

print(pd.concat(dfs, axis=1).fillna(""))

打印：

  England_Premier League    England_Division 1 England_Division 2
0             Harry Kane            Will Grigg         Eoin Doyle
1          Mohamed Salah  Jonson Clarke-Harris       Matt Watters
2        Kevin De Bruyne           Jerry Yates       James Vughan
3                                   Ivan Toney                   
4                                 Troy Parrott                   

【讨论】：

很好的答案

@andrej 看起来很完美，如果您不介意，我只是有几个问题。您是否将原始文本文件称为“txt”？如果不是太麻烦的话，您是否可以输入一些简单的 cmets，以便我可以确切地看到每一行代码在做什么？

@PythonBeginner txt 只是变量名。您可以加载文件，例如 txt = open("your_file.txt", "r").read()

@AndrejKesely 谢谢。我现在必须对 For 循环进行一些研究，看看它是如何提取数据的。

@AndrejKesely 你介意带我了解一下使用的 r.compile 方法吗