Question

我有几个csv格式的文件，例如100-age.csv 100-rel.csv 100-gender.csv 101-age.csv ... 101-gender.csv ... 482-rel.csv 482-gender.csv等。我必须为每个索引创建新文件，即100-combo.csv，它将合并100-age.csv 100-rel.csv和100-gender.csv 水平。我可以使用熊猫对一个文件执行此操作

import pandas as pd

age = pd.read_csv('100-age.csv', header=None)
gender = pd.read_csv('100-gender.csv', header=None)
rel = pd.read_csv('100-rel.csv', header=None)

combined = pd.concat([age, gender, rel], axis=1)

combined.to_csv('100-combo.csv', header=None, index=None)

使用Linux，有cat之类的方法只能垂直添加，即彼此堆叠，而paste的命令会干扰我在这些文件中的格式。

    def merged_data(i):   

        age = pd.read_csv(path+str(i)+'.pdf-age.csv', header=None, error_bad_lines=False)
        gender = pd.read_csv(path+str(i)+'.pdf-gender.csv', header=None, error_bad_lines=False)
        rel = pd.read_csv(path+str(i)+'.pdf-rel.csv', header=None, error_bad_lines=False)
        combined = pd.concat([age, gender, rel], axis=1)
        combined['block'] = str(i)
        combined.to_csv(path+str(i)+'-combo.csv', header=None, index=None)

for num in range(1,483):

    merged_data(num)

我收到此错误

EmptyDataError: No columns to parse from file

但是，我知道我所有的数据文件都有一些或其他值

Answer 1

我做到了，得到了我想要的。我用

import pandas as pd
import numpy as np
from pandas.io.common import EmptyDataError

    def merged_data(i):   
        try:
            age = pd.read_csv(path+str(i)+'.pdf-age.csv', header=None, error_bad_lines=False, delim_whitespace=True)
        except EmptyDataError:
            age = pd.DataFrame()
        try:  
            gender = pd.read_csv(path+str(i)+'.pdf-gender.csv', header=None, error_bad_lines=False, delim_whitespace=True)
        except EmptyDataError:
            gender = pd.DataFrame()
        try:
            rel = pd.read_csv(path+str(i)+'.pdf-rel.csv', header=None, error_bad_lines=False, delim_whitespace=True)
        except EmptyDataError:
            rel = pd.DataFrame()
            combined = pd.concat([age, gender, rel], axis=1)
            combined['block'] = str(i)
            combined.to_csv(path+str(i)+'-combo.csv', header=None, index=None)

for num in range(1,483):

    merged_data(num)

读取多个CSV文件并将其水平合并

1 个答案: