Python新手跟随教程操作时遇DataFrame无to_frame属性报错求助

Hey there! Let's work through this error step by step.

First, let's recap your code logic for calculating missing value ratios:

miss = train.isnull().sum()/len(train)
miss = miss[miss>0]
miss.sort_values(inplace = True)

This code should return a pandas Series (not a DataFrame) where each index is a column name from your train DataFrame, and the values are the corresponding missing value ratios.

The error 'DataFrame' object has no attribute 'to_frame' tells us that when you tried calling .to_frame(), you were actually calling it on a DataFrame instead of the miss Series you intended. Here are the most likely reasons and fixes:

1. You accidentally overwrote the miss variable #

Check if you have any code after the three lines above that reassigns miss to a DataFrame. For example, something like:

# This would turn miss into a DataFrame and overwrite your original Series
miss = train[['Electrical', 'MasVnrType']]

If that's the case, rename your DataFrame variable to something else (like selected_cols) to keep the original miss Series intact.

2. You called .to_frame() on the wrong variable #

Double-check that you're running miss.to_frame() and not train.to_frame() (or another DataFrame variable). to_frame() is a method exclusive to pandas Series, which is why calling it on a DataFrame throws an error.

3. Verify the type of miss #

To confirm what type of object miss is, add this line right before you call .to_frame():

print(type(miss))

• If it outputs <class 'pandas.core.series.Series'>, then miss.to_frame() will work perfectly. You can even specify a column name for the resulting DataFrame:

  miss_df = miss.to_frame(name='missing_ratio')
  

• If it outputs <class 'pandas.core.frame.DataFrame'>, you'll need to trace back your code to find where miss got converted to a DataFrame.

As a side note, since you're working on a machine learning project, once you have your missing value ratios in a DataFrame, you can easily visualize them with:

miss_df.plot(kind='barh', figsize=(10,6))

内容的提问来源于stack exchange，提问作者Serdia