如何将 a.fillna() 应用于过滤后的 dataframe？

Question

I am learning pandas using a dataset of players from the womans world cup.

I want to apply a.fillna() on the results of the following filter of the main dataframe:

df[(df["pos"] == "GK") & (df["goals"].isnull())]

How can I apply:

fillna(0, inplace=True)

To the above filter, , on the 'goals' column.

I did use the below code but when I apply it, it runs but doesn't seem to work:

df.loc[df["pos"] == "GK"].loc[df["goals"].isnull()].loc[:,"goals"].fillna(0, inplace=True)

Answer 1

The first line selects the index and you know the column, so just use one loc and it should be fine:

df.loc[(df["pos"] == "GK") & (df["goals"].isnull()), 'goals'].fillna(0, inplace=True)

update:

So it seems pandas returns a copy and inplace doesn't really do anything there. However, you don't want to assign it to all of your dataframe. Since you are already filtering for nan values, an easier way is to just assign the value directly. df.loc[(df["pos"] == "GK") & (df["goals"].isnull()), 'goals'] = 0 This worked fine on my computer, hope it solves your problem.

Answer 2

df[(df["pos"] == "GK") & (df["goals"].isnull()), 'goals']=0

That will solve your problem.

Answer 3

You already masked using a boolean select. So just .fillna(0)

df[(df["pos"] == "GK") & (df["goals"].isnull())].fillna(0)