如何优化此替换列和索引的代码？答案

【问题标题】：How can I optimize this code that does replacement of columns and indexing?如何优化此替换列和索引的代码？
【发布时间】：2017-12-05 11:42:33
【问题描述】：

census_subdivision_profile_merged 是一个数据框，我在这里分别做很多事情。有没有办法一次完成所有这些？

# Drop missing data
census_subdivision_profile_merged = census_subdivision_profile_merged.dropna()
census_subdivision_profile_merged.columns = [census_subdivision_profile_merged.columns[i]+' '+str(i) for i in range(len(census_subdivision_profile_merged.columns))]
census_subdivision_profile_merged.columns = [census_subdivision_profile_merged.columns[i].replace(" ", "_") for i in range(len(census_subdivision_profile_merged.columns))]
census_subdivision_profile_merged.columns = [census_subdivision_profile_merged.columns[i].replace(",", "_") for i in range(len(census_subdivision_profile_merged.columns))]
census_subdivision_profile_merged.columns = [census_subdivision_profile_merged.columns[i].replace("-", "_") for i in range(len(census_subdivision_profile_merged.columns))]
census_subdivision_profile_merged.columns = [census_subdivision_profile_merged.columns[i].replace("%", "_") for i in range(len(census_subdivision_profile_merged.columns))]
census_subdivision_profile_merged.columns = [census_subdivision_profile_merged.columns[i].replace("$", "_") for i in range(len(census_subdivision_profile_merged.columns))]

【问题讨论】：

您是否只清理列标题中的文本？你能发布一个小的示例数据框吗？

标签： python regex pandas dataframe data-cleaning

【解决方案1】：

您使用字符串执行 5 次替换方法，但您也可以使用正则表达式，例如：

import re

#Test data frame
df=pd.DataFrame({"data1":["E %,-$p,e","E    $m$$-%ple"],"data2":["E %,-$p,e","E    $m$$-%ple"]})



#Remove all special characters and whitespaces for each row, for each word
for j in df.columns:
    for strs in range(len(df[j])):
        df.loc[strs,j]=re.sub(r'[-%,$\s]',"_",df.loc[strs,j])


print(df)

对于您的示例应该可以：

for  j in census_subdivision_profile_merged.columns:
    for strs in range(len(census_subdivision_profile_merged[j])):
        census_subdivision_profile_merged.loc[strs, j] = re.sub(r'[-%,$\s]', "_", census_subdivision_profile_merged.loc[strs, j])

或者，您可以尝试从完整的 DataFrame 中删除所有特殊字符和空格。

census_subdivision_profile_merged = census_subdivision_profile_merged.replace(r"[-%,$\s]","_",regex=True)

【讨论】：