通过基于另一列的模式估算列值

Question

I have a dataset that contains n columns. Those include a column for the Birth City and another for the Birth Country. What I want to do is, based on the Birth Country, I am going to get the mode of the Birth City, and use it for the missing values in the Birth City column.

I tried the following code, but nothing is affected.

df["Birth City"]= df.groupby('Birth Country')['Birth City'].transform(lambda x: x.fillna(x.mode()))
df[df["Birth City"].isnull()]

After executing the above code, I still get the same missing Birth City values.

Answer 1

You need to replace x.fillna(x.mode()) with x.mode()[0] in your code

df["Birth City"]= df.groupby('Birth Country')['Birth City'].transform(lambda x: x.mode()[0])
df[df["Birth City"].isnull()]

In your example you are trying to use Series.transform which applies lambda x: x.fillna(x.mode()) . This lambda function replaces x with x.fillna(x.mode()) (not x.mode() ) which equals to Series with missing values filled. x.mode() also has type Series and equals to Series of modes in sorted order, so you have to use x.mode()[0]