BUG: Resampling PeriodIndex-ed to multiple of frequencies not working as expected

[winklerand]

Code Sample, a copy-pastable example if possible

In [1]: import numpy as np

In [2]: import pandas as pd


In [3]: s = pd.Series([2017, 2018], index=pd.period_range('2017', freq='A', periods=2))


In [4]: s.resample('2Q', kind='period').ffill()
Warning: multiple of frequency -> timestamps
Out[4]: 
2017-03-31    2017
2017-09-30    2017
2018-03-31    2018
Freq: 2Q-DEC, dtype: int64

In [7]: s2 = pd.Series(np.arange(12), index=pd.period_range('2017-01', freq='M', periods=12))


In [8]: s2.resample('2Q', kind='period').mean()
Warning: multiple of frequency -> timestamps
Out[8]: 
2017-03-31     1.0
2017-09-30     5.5
2018-03-31    10.0
Freq: 2Q-DEC, dtype: float64

To compare with, results for resampling to base frequency (no multiples) returning PeriodIndex-ed and covering the full original time span:


In [5]: s.resample('Q', kind='period').ffill()
Out[5]: 
2017Q1    2017
2017Q2    2017
2017Q3    2017
2017Q4    2017
2018Q1    2018
2018Q2    2018
2018Q3    2018
2018Q4    2018
Freq: Q-DEC, dtype: int64

In [9]: s2.resample('Q', kind='period').mean()
Out[9]: 
2017Q1     1
2017Q2     4
2017Q3     7
2017Q4    10
Freq: Q-DEC, dtype: int64

Problem description

• I'd expect resampling PeriodIndex-ed series/dataframes would return a PeriodIndex-ed result by default, even more when givenkind='period'.
• Moreover, I'd expect the original full time span to be covered by the resampling result (upsampling A->2Q would return 2 periods per year, downsampling M->2Q would return 2 periods per 12 months).

As indicated by the warning message, resampling to multiple of frequencies falls back to timestamp-based resampling. Both work fine when resampling to a "base" frequency without any multiple.

Expected Output

In [4]: s.resample('2Q', kind='period').ffill()
Out[4]:
2017Q1    2017
2017Q3    2017
2018Q1    2018
2018Q3    2018
Freq: 2Q-DEC, dtype: int64

In [8]: s2.resample('2Q', kind='period').mean()
Out[8]:
2017Q1    2.5
2017Q3    8.5
Freq: 2Q-DEC, dtype: float64

Output of pd.show_versions()

INSTALLED VERSIONS ------------------ commit: None python: 3.6.1.final.0 python-bits: 64 OS: Darwin OS-release: 15.6.0 machine: x86_64 processor: i386 byteorder: little LC_ALL: None LANG: de_DE.UTF-8 LOCALE: de_DE.UTF-8

pandas: 0.19.2
nose: None
pip: 9.0.1
setuptools: 33.1.1.post20170320
Cython: None
numpy: 1.12.1
scipy: 0.19.0
statsmodels: 0.8.0
xarray: None
IPython: 5.3.0
sphinx: None
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2016.10
blosc: None
bottleneck: 1.2.0
tables: None
numexpr: 2.6.2
matplotlib: 2.0.0
openpyxl: None
xlrd: None
xlwt: None
xlsxwriter: None
lxml: None
bs4: 4.5.3
html5lib: 0.999
httplib2: None
apiclient: None
sqlalchemy: None
pymysql: None
psycopg2: None
jinja2: 2.9.5
boto: None
pandas_datareader: None

[jreback]

xref #12884

there are quite a few issues w.r.t. PeriodIndex and resampling. just need someone motivated to digin an fix.

this is basically a dupe as things are returned as a DTI. But i'll link this for tests and such.