pandas data with double quote

问题

I am trying to read a large dataset in .csv format which is update automatically using the pandas library. The problem is that in my data, the first row is a string without double quotation marks, and the other colums are strings with double quotation marks. It is not possible for me to adjust the .csv file manually.

A simplified dataset would look like this

1. A,"B","C","D"
2. comp_a,"tree","house","door"
3. comp_b,"truck","red","blue"

I need the data to be stored as separate columns without the quotation marks like this:

1. A B C D
2. comp_a tree house door
3. comp_b truck red blue

I tried using

import pandas as pd
df_csv = pd.read(path_to_file,delimiter=',')

which gives me the complete header as a single variable for the last column

1. A,"B","C","D"
2. comp_a "tree" "house" "door"
3. comp_b "truck" "red" "blue"

The closest result to the one i need was by using the following

df_csv = pd.read(path_to_file,delimiter=',',quoting=3)

which correctly recognizes each column, but adds in a bunch of extra double quotes.

1. "A ""B"" ""C"" ""D"""
2. "comp_a ""tree"" ""house"" ""door"""
3. "comp_b ""truck"" ""red"" ""blue"""

Setting quoting to a value from 0 to 2 just reads an entire row as a single column.

Does anyone know how I can remove all quotation marks when reading the .csv file?

回答1:

Just load the data with pd.read_csv() and then use .replace('"','', regex=True)

In one line it would be:

df = pd.read_csv(filename, sep=',').replace('"','', regex=True)

To set the columns names:

df.columns = df.iloc[0]

And drop row 0:

df = df.drop(index=0).reset_index(drop=True)

回答2:

you can replace " after read_csv and save that file again using df_csv.to_csv('fname')

df_csv.apply(lambda x:x.str.replace('"', ""))

回答3:

Consider your data in a file data.csv like

$> more data.csv 
A,"B","C","D"
comp_a,"tree","house","door"
comp_b,"truck","red","blue"

Perhaps a newer pandas version would solve your problem from itself, e.g. at pd.__version__ = '0.23.1'

In [1]: import pandas as pd

In [2]: pd.read_csv('data.csv')
Out[2]: 
        A      B      C     D
0  comp_a   tree  house  door
1  comp_b  truck    red  blue

Otherwise apply a replace on the read-out

pd.read_csv('data.csv').replace('"', '')

来源：https://stackoverflow.com/questions/51359010/pandas-data-with-double-quote