从Python目录中的多个CSV文件中提取特定列

编程问答 2022-06-10

问题描述

我在一个目录中大约有200个CSV文件，其中包含不同的列，但是有些文件中有我要提取的数据。我要拉的一列称为“ Programme”（行的顺序不同，但名称相同），另一列包含“ would荐”（并非所有措辞都相同，但它们都将包含该措辞）。最终，我想为每个CSV提取这些列下的所有行，并将它们附加到仅包含这两列的数据框中。我尝试过仅使用一个CSV来完成此操作，但无法使其正常工作。这是我尝试过的：

import pandas as pd
from io import StringIO

df =  pd.read_csv("test.csv")

dfout = pd.DataFrame(columns=['Programme','Recommends'])

for file in [df]:
    dfn = pd.read_csv(file)
    matching = [s for s in dfn.columns if "would recommend" in s]
    if matching:
        dfn = dfn.rename(columns={matching[0]:'Recommends'})
        dfout = pd.concat([dfout,dfn],join="inner")

print(dfout)

我收到以下错误消息，所以我认为这是一个格式问题（它不喜欢pandas df？）： ValueError（msg.format（_type = type（filepath_or_buffer））） ValueError：无效的文件路径或缓冲区对象类型：

当我尝试这样做时：

csv1 = StringIO("""Programme,"Overall,I am satisfied with the quality of the programme",I would recommend the company to a friend or colleague,Please comment on any positive aspects of your experience of this programme
Nursing,4,IMAGE
Nursing,1,3,very good
Nursing,5,I enjoyed studying tis programme""")

csv2 = StringIO("""Programme,I would recommend the company to a friend,The programme was well organised and running smoothly,It is clear how students' Feedback on the programme has been acted on
IT,2,4
IT,5
IT,5""")

dfout = pd.DataFrame(columns=['Programme','Recommends'])

for file in [csv1,csv2]:
    dfn = pd.read_csv(file)
    matching = [s for s in dfn.columns if "would recommend" in s]
    if matching:
        dfn = dfn.rename(columns={matching[0]:'Recommends'})
        dfout = pd.concat([dfout,join="inner")

print(dfout)

这正常工作，但我需要读取CSV文件。有什么想法吗？

以上示例的预期输出：

enter image description here

解决方法

以下作品：

import pandas as pd
import glob

dfOut = []

for myfile in glob.glob("*.csv"):
    tmp = pd.read_csv(myfile,encoding='latin-1')
    
    matching = [s for s in tmp.columns if "would recommend" in s]
    if len(matching) > 0:
        tmp.rename(columns={matching[0]: 'Recommend'},inplace=True)
        tmp = tmp[['Subunit','Recommend']]
        dfOut.append(tmp)
        
df = pd.concat(dfOut)
csvcsvcsvmatchingpandaspandaspython