使用带有正则表达式分隔符的 read_csv 读取文本文件

Question

我一直在尝试使用 read_csv 读取文本文件 university_towns.txt，但如屏幕截图所示，在使用正则表达式分隔符解析文件时，如下代码所示，我收到以下错误：

ParserError: Expected 2 fields in line 89, see 3. 错误可能 是因为使用多字符分隔符时引号被忽略。

有什么办法可以解决这个问题，似乎只有一个地方的双引号有问题，还请解释为什么会这样？我也尝试使用 quotechar 参数，但不明白如何使用它。

我读取文件的代码如下：

university_towns = pd.read_csv('university_towns.txt', sep= "\s\(", engine='python', header=None)

university_towns.txt file image

Annville (Lebanon Valley College)[2]
Bethlehem (Lehigh University, Moravian College)
Bloomsburg (Bloomsburg University of Pennsylvania)[2]
Bradford (University of Pittsburgh at Bradford)
California (California University of Pennsylvania)[2]
Carlisle (Dickinson College)
Cecil B. Moore, Philadelphia, also known as "Templetown" (Temple University)
Clarion (Clarion University of Pennsylvania)[2]
Collegeville (Ursinus College)
Cresson (Mount Aloysius College)[2]
East Stroudsburg (East Stroudsburg University of Pennsylvania)[2]
Edinboro (Edinboro University of Pennsylvania)[2]
Erie (Gannon University, Mercyhurst College, Penn State Erie)
Gettysburg (Gettysburg College)[2]
Greensburg (Seton Hill University, University of Pittsburgh at Greensburg)
Grove City (Grove City College)[2]
Huntingdon (Juniata College)[2]
Indiana (Indiana University of Pennsylvania)[2]
Johnstown (University of Pittsburgh at Johnstown)
Kutztown (Kutztown University of Pennsylvania)[2]
Lancaster (Franklin & Marshall)
Carrollton (University of West Georgia)[2]*Dahlonega (North Georgia College & State University)[2]

我在上面粘贴了一些文本文件的行。另外，最后一行是第 89 行。

Answer 1

您的文本样本不会重现该错误。您可以通过在调用中添加一些参数来提示导致错误的原因：

pd.read_csv('university_towns.txt', sep= "\s\(", 
            engine='python', header=None , 
            error_bad_lines= False, warn_bad_lines = True)

见pandas.read_csv()

---

使用

# https://pandas.pydata.org/pandas-docs/stable/generated/pandas.read_csv.html

import pandas as pd

with open( 'university_towns.txt', "w") as f:
    f.write("""Annville (Lebanon Valley College)[2]
Bethlehem (Lehigh University, Moravian College)
Bloomsburg (Bloomsburg University of Pennsylvania)[2]
Carrollton (University of West Georgia)[2]*Dahlonega (North Georgia College & State University)[2]""")

university_towns = pd.read_csv('university_towns.txt', sep= "\s\(", engine='python', header=None)

print(university_towns)

确实会重现您的错误。原因是最后一行（看起来应该是单独的行）包含 2 个匹配您的正则表达式，因此希望分成 3 列，而所有其他行只有 2 列。 => 错误。

要修复它，请将违规行拆分为 2：

Annville (Lebanon Valley College)[2]
Bethlehem (Lehigh University, Moravian College)
Bloomsburg (Bloomsburg University of Pennsylvania)[2]
Carrollton (University of West Georgia)[2]
Dahlonega (North Georgia College & State University)[2]

它会起作用的。

Answer 2

显示第 83 列，这里可以看到其他一些 - 我认为那里有两个 ( - \s\(。至少这就是此错误消息的含义。其他可能的问题是那里有奇怪的字符，解析器根本就丢失了。我认为大学名称不可能......无论如何 - 看看那条线。如果这不明显，请与我们分享。