Question

I have a Pandas DataFrame in the following format.

I am trying to fill the NaN value by using the most recent non-NaN value and adding one second to the time value. For example, in this case, the program should take the most recent non-NaN value of 8:30:20 and add one second to replace the NaN value. So, the replacement value should be 8:30:21. Is there a way in Pandas to simulate this process for the entire column?

Answer 1

Considering a numerical dataset, you can use the following function:

def fill(df, column):
    cache = None
    for index, value in enumerate(df[column]):
        if pd.isna(value):
            if(cache):
                df.loc[index, column] = cache + 1
                cache = cache + 1
        else:
            cache = value
    return df

There are two special cases you did not consider in your description:

What if the first value of the column is None/NaN ? in my implementation, this value remains as None/NaN given that there is no "most recent value". You can pass a parameter to indicate a value for this case.
What if there are two None/NaN in a row ? in my implementation, the most recent value for the second None/NaN is going to be the new value of the first None/NaN. This behaviour is defined in the 7th line.

友情链接