I have a time series as pandas dataframe, that have one column named "pre". The index of the dataframe is pandas date_range. The dates range from 1998-01-01 to 2002-12-31.
I want to get the sum of pre values from "1998-01-01" to "1998-05-31" plus "1998-09-01" to "1998-12-31". After getting this for the year 1998, how can I get the same for all the years?
Use dictionary comprehension with f-strings:
np.random.seed(2020)
r = pd.date_range('1998-01-01','2002-12-31')
s = pd.Series(np.random.randint(10, size=len(r)), index=r)
# print (s)
out = {y: s.loc[f"{y}-01-01":f"{y}-05-31"].sum() + s.loc[f"{y}-09-01":f"{y}-12-31"].sum()
for y in range(1998, 2003)}
print (out)
{1998: 1235, 1999: 1201, 2000: 1154, 2001: 1270, 2002: 1184}
Another solution is convert DatetimeIndex to months and remove 6,7,8 months and then grouping by years with aggregate sum:
s1 = s[~s.index.month.isin([6,7,8])]
print (s1.groupby(s1.index.year).sum())
1998 1235
1999 1201
2000 1154
2001 1270
2002 1184
dtype: int32
