I have a large pandas dataframe and am struggling to set NaN values of specific columns to 1. The column types for the columns I want to work on are below:

guests = object, beds = float64, bathrooms = float64, bedrooms = object.

I have tried these methods but none have worked:

df['guests', 'bedrooms', 'beds', 'bathrooms'] = df['guests', 'bedrooms', 'beds', 'bathrooms'].replace(np.nan, 1)

df['guests', 'bedrooms', 'beds', 'bathrooms'].fillna(1, inplace=True)

Any help appreciated.

CodePudding user response：

Don't use inplace as you are modifying a copy of the DataFrame, but assign back or update:

df.update(df[['guests', 'bedrooms', 'beds', 'bathrooms']].fillna(1))

Or:

cols = ['guests', 'bedrooms', 'beds', 'bathrooms']
df[cols] = df[cols].fillna(1)

Or limit the values in fillna using a dictionary:

cols = ['guests', 'bedrooms', 'beds', 'bathrooms']
df.fillna({k: 1 for k in cols}, inplace=True)

CodePudding user response：

Use DataFrame.fillna with convert columns names by dictionary to 1:

df = df.fillna(dict.fromkeys(['guests', 'bedrooms', 'beds', 'bathrooms'], 1))