如何用带有逗号分隔符和空格的pandas解析csv？

Question

我目前有以下data.csv，它有逗号分隔符：

name,day
Chicken Sandwich,Wednesday
Pesto Pasta,Thursday
Lettuce, Tomato & Onion Sandwich,Friday
Lettuce, Tomato & Onion Pita,Friday
Soup,Saturday

解析器脚本是：

import pandas as pd


df = pd.read_csv('data.csv', delimiter=',', error_bad_lines=False, index_col=False)
print(df.head(5))

输出是：

Skipping line 4: expected 2 fields, saw 3
Skipping line 5: expected 2 fields, saw 3

               name        day
0  Chicken Sandwich  Wednesday
1       Pesto Pasta   Thursday
2              Soup   Saturday

我如何处理案件Lettuce, Tomato & Onion Sandwich。每个项目应该用,分隔，但是一个项目可能有一个逗号，后跟一个空格。所需的输出是：

                               name        day
0                  Chicken Sandwich  Wednesday
1                       Pesto Pasta   Thursday
2  Lettuce, Tomato & Onion Sandwich     Friday
3      Lettuce, Tomato & Onion Pita     Friday
4                              Soup   Saturday

Answer 1

另一种适用于其他情况的替代方案。好的，这很难看。

import pandas as pd
from io import StringIO

for_pd = StringIO()
with open('theirry.csv') as input:
    for line in input:
        line = line.rstrip().replace(', ', '|||').replace(',', '```').replace('|||', ', ').replace('```', '|')
        print (line, file=for_pd)
for_pd.seek(0)

df = pd.read_csv(for_pd, sep='|')

print (df)

结果：

                               name        day
0                  Chicken Sandwich  Wednesday
1                       Pesto Pasta   Thursday
2  Lettuce, Tomato & Onion Sandwich     Friday
3      Lettuce, Tomato & Onion Pita     Friday
4                              Soup   Saturday

Answer 2

这可能有所帮助。

import pandas as pd
p = "PATH_TO.csv"
df = pd.read_csv(p, delimiter='(,(?=\S)|:)')
#print(df.head(5))
print "-----"
print df["name"]
print "-----"
print df["day"]

输出：

-----
0                    Chicken Sandwich
1                         Pesto Pasta
2    Lettuce, Tomato & Onion Sandwich
3        Lettuce, Tomato & Onion Pita
4                                Soup
Name: name, dtype: object
-----
0    Wednesday
1     Thursday
2       Friday
3       Friday
4     Saturday
Name: day, dtype: object