Python中DataFrame object类型列相乘添加新列的警告解决

问题背景 #

这是Udemy训练营课程（第76天）的问题，我是Python新手，若表述不当请见谅。我有一个包含10列的apps.csv文件，各列数据类型如下：

App                object
Category           object
Rating            float64
Reviews             int64
Size_MBs          float64
Installs           object
Type               object
Price              object
Content_Rating     object
Genres             object
dtype: object

需求与现有代码 #

我希望将文件中的Price和Installs列相乘，添加名为Revenue_Estimate的新列，编写代码如下：

import pandas as pd
df=pd.read_csv('apps.csv')
clean_df=df.dropna()
clean_df.loc[:, "Installs"] = clean_df.loc[:, "Installs"].astype(str).str.replace(",", "")
clean_df.loc[:, "Installs"] = pd.to_numeric(clean_df.loc[:, "Installs"])
clean_df.loc[:,'Price'] = clean_df.loc[:,'Price'].astype(str).str.replace('$', "")
clean_df.loc[:, "Price"] = pd.to_numeric(clean_df.loc[:,'Price'])
#Line 8:  
clean_df['Revenue_Estimate']= clean_df.Installs * clean_df.Price

遇到的警告 #

执行上述代码时出现如下警告：

A value is trying to be set on a copy of a slice from a DataFrame. Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation:
 clean_df['Revenue_Estimate']= clean_df.Installs * clean_df.Price

我也尝试了以下代码，仍无法解决：

df_apps_clean['Revenue_Estimate'] = df_apps_clean.Installs.mul(df_apps_clean.Price)

对应的警告信息：

A value is trying to be set on a copy of a slice from a DataFrame. Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation:
clean_df['Revenue_Estimate'] = clean_df.Installs.mul(clean_df.Price)

解决方案 #

问题出在clean_df = df.dropna()这一行：dropna()默认返回的是原DataFrame的视图而非独立副本，后续对clean_df的修改操作会触发Pandas的视图/副本警告。

解决方法很简单，只需要在dropna()时显式指定返回副本，或者直接对结果调用.copy()：

修改后的完整代码 #

import pandas as pd
df=pd.read_csv('apps.csv')
# 方法1：dropna时指定copy=True
clean_df=df.dropna(copy=True)
# 或者方法2：显式复制
# clean_df=df.dropna().copy()

clean_df.loc[:, "Installs"] = clean_df.loc[:, "Installs"].astype(str).str.replace(",", "")
clean_df.loc[:, "Installs"] = pd.to_numeric(clean_df.loc[:, "Installs"])
clean_df.loc[:,'Price'] = clean_df.loc[:,'Price'].astype(str).str.replace('$', "")
clean_df.loc[:, "Price"] = pd.to_numeric(clean_df.loc[:,'Price'])

# 现在添加新列不会再触发警告
clean_df['Revenue_Estimate']= clean_df.Installs * clean_df.Price

这样修改后，clean_df就变成了独立的DataFrame副本，后续的所有修改操作都不会再触发视图/副本相关的警告。

内容的提问来源于stack exchange，提问作者Venkat Karthik