Pandas Dataframe 复杂计算

Question

I have the following dataframe,df:

     Year  totalPubs  ActualCitations
0   1994         71       191.002034
1   1995         77      2763.911781
2   1996         69      2022.374474
3   1997         78      3393.094951

I want to write code that would do the following:

Citations of currentyear / Sum of totalPubs of the two previous years

I want something to create a new column called Impact Factor, and generate it as follows:

for index, row in df.iterrows():
    if row[0]>=1996:
        df.at[index,'Impact Factor'] = df.at[index, 'ActualCitations'] / (df.at[index-1, 'totalPubs'] + df.at[index-2, 'totalPubs'])

Answer 1

I believe the following does what you want:

In [24]:
df['New_Col'] = df['ActualCitations']/pd.rolling_sum(df['totalPubs'].shift(), window=2)
df

Out[24]:
   Year  totalPubs  ActualCitations    New_Col
0  1994         71       191.002034        NaN
1  1995         77      2763.911781        NaN
2  1996         69      2022.374474  13.664692
3  1997         78      3393.094951  23.240376

So the above uses rolling_sum and shift to generate the previous 2 years sum and we then divide the citations value by that value.