Python – Pandas groupby : group by semester

I need to group the data by semester, but there is no frequency label available here

2QS (2 quarters from start) and 6MS (6 months from start) won’t do because they will start at different moments, according to the first datetime of my data frame. (Very counterintuitive and error-prone, IMHO: I didn’t see this issue until I used a different dataset from May instead of January…)

).

from datetime import *
import pandas as pd
import numpy as np

df = pd. DataFrame()

days = pd.date_range(start="2017-05-17", 
                     end="2017-11-29",
                    freq="1D")
df = pd. DataFrame({'DTIME': days, 'DATA': np.random.randint(50, high=80, size=len(days))})
df.set_index('DTIME', inplace=True)

grouped = df.groupby(pd. Grouper(freq='2QS'))
print("Groups date start:")
for dtime, group in grouped:
    print dtime
    # print(group)

Return

Groups date start:
2017-04-01 00:00:00   <== because my first datetime is in May, 2017
2017-10-01 00:00:00

Instead of:

Groups date start:
2017-01-01 00:00:00   <== I want the semesters referred to the year!
2017-06-01 00:00:00

As a possible workaround, I created two new columns in my data frame and then grouped based on them:

      df["year"] = df.index.year.astype(int)
      df["semester"] = df.index.month.astype(int)
      df["semester"] = df["semester"] - 1
      df["semester"] = df["semester"] // 6
      grouped = df.groupby(["year", "semester"])

Is this the only way?

There are two other minor issues that, just out of curiosity, don’t deserve a separate stackoverflow issue:

1. Why is the label W (weekend) available, but WS (start of week) not?

2. How to write this in one line?

     df["semester"] = df.index.month.astype(int)
     df["semester"] = df["semester"] - 1
     df["semester"] = df["semester"] // 6
   

The closest is >anchored-offsets , but a month gone.

The second :

df["semester"] =  (df.index.month.astype(int) - 1) // 6

Or do not create a new column:

years = df.index.year.astype(int)
semes = (df.index.month.astype(int) - 1) // 6
grouped = df.groupby([years, semes])