如何将UDF应用于数据框？

Question

I am trying to create a function that will cleanup and dataframe that I put through the function. But I noticed that the df returned is cleanued up but not in place of the original df.

How can I run a UDF on a dataframe and keep the updated dataframe saved in place?

ps I know I can combine these rules into one line but the function I am creating is a lot more complex so I don't want to combine for this example

df = pd.DataFrame({'Key': ['3', '9', '9', '9', '9','34','34', '34'], 
              'LastFour': ['2290', '0087', 'M433','M433','25','25','25','25'],
              'NUM': [20120528, 20120507, 20120615,20120629,20120621,20120305,20120506,20120506]})

def cleaner(x):
    x = x[x['Key'] == '9']
    x = x[x['LastFour'] == 'M433']
    x = x[x['NUM'] == 20120615]
    return x

cleaner(df)

Result from the UDF:

    Key LastFour    NUM
2   9   M433        20120615

But if I run the df after the function then I still get the original dataset:

    Key LastFour   NUM
0   3   2290       20120528
1   9   0087       20120507
2   9   M433       20120615
3   9   M433       20120629
4   9   25         20120621
5   34  25         20120305
6   34  25         20120506
7   34  25         20120506

Answer 1

You need to assign the result of cleaner(df) back to df as so:

df = cleaner(df)

An alternative method is to use pd.DataFrame.pipe to pass your dataframe through a function:

df = df.pipe(cleaner)