Pandas - 获取 pandas 中一行的索引应用 function

Question

Here is my pandas code:

def calcObj(row):
    d = dict(calc1 = iferror(row.Hours1.sum(), row.Hours2.sum(), 0, '+'))
    
    if row.Process == 'A': # this doesn't work
        d['ProcessKey'] = 700
    else:
        d['ProcessKey'] = 500
    
    return pd.Series(d)

    
df.groupby(['MainProcess']).apply(calcObj)

I am trying to check if a process name is A and if it is return a different value. Unfortunately it doesn't work and i get the following error:

AttributeError: 'DataFrame' object has no attribute 'Process '

I assume it's because i am not grouping by process only by MainProcess.

Is there any way to get access to this item within the apply function? Any other work-around would also be very helful

Here is my example dataframe, BG/MainProcess, CoreProcess and Process1 are indexes, Hours1/Hours2 are columns:

Bg              MainProcess                CoreProcess          Process      Hours1      Hours2                                        
Building1       MainProcess-1              CoreProcess-1        S-Process-1     150         250
                                                                S-Process-2     150         250
                                           CoreProcess-2        S-Process-3     150         250
                                                                S-Process-1     150         250
                                                                S-Process-2     150         250
Building2       MainProcess-2              CoreProcess-3        S-Process-1     150         250
                                                                S-Process-2     150         250
                MainProcess-3              CoreProcess-4        S-Process-1     150         250
                                                                S-Process-2     150         250
                                                                S-Process-3     150         250

Answer 1

Beware, the columns in the index are not columns of the DataFrame!

In your example, as Process is in the (multi-) index, df['Process'] will raise a KeyError independently of the groupby . You must reset the column from the index to be able to use it. For example you could reset it before the groupby:

df.reset_index(level='Process').groupby(['MainProcess']).apply(calcObj)

But beware : calcObj will not receive rows here but the sub-dataframes having same values in the MainProcess column...