loffset in resample does not work with ffill, bfill, pad

[badge]

Minimal example

df = pd.DataFrame(np.random.rand(49),
                  index=pd.DatetimeIndex(start='2018-04-01',
                                         end='2018-04-03',
                                         freq='1h')
                 )

print(df.head())
print(df.resample('D', loffset=pd.Timedelta('12h')).mean())
print(df.resample('D', loffset=pd.Timedelta('12h')).ffill())

Problem description

DataFrame and DataFrameGroupBy both provide a resample method, for resampling timeseries data. These methods have an optional argument, loffset, which accepts a pandas Timedelta instance and is used to adjust the output DataFrame's index, once a suitable aggregation method has been used on the resulting DatetimeIndexResampler or DatetimeIndexResamplerGroupby object.

This works as expected for, for example, mean, but fails in the cases of ffill, bfill and pad. The resample is successful, but the loffset argument is ignored.

Expected Output

                        0

2018-04-01 00:00:00 0.246580
2018-04-01 01:00:00 0.380124
2018-04-01 02:00:00 0.227830
2018-04-01 03:00:00 0.376657
2018-04-01 04:00:00 0.139719
0
2018-04-01 12:00:00 0.434068
2018-04-02 12:00:00 0.471917
2018-04-03 12:00:00 0.395527
0
2018-04-01 12:00:00 0.246580
2018-04-02 12:00:00 0.402313
2018-04-03 12:00:00 0.395527

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.4.final.0
python-bits: 64
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 61 Stepping 4, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None

pandas: 0.22.0
pytest: 3.0.7
pip: 9.0.3
setuptools: 39.0.1
Cython: 0.27.3
numpy: 1.14.2
scipy: 1.0.0
pyarrow: None
xarray: None
IPython: 6.3.1
sphinx: 1.7.2
patsy: 0.4.1
dateutil: 2.7.2
pytz: 2018.4
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.2
feather: None
matplotlib: 2.1.2
openpyxl: 2.4.7
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.3
bs4: 4.6.0
html5lib: 1.0.1
sqlalchemy: 1.1.9
pymysql: None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None

[jreback]

hmm, guess it doesn. a pull-request would be great to fix! (also doesn't work for .asfreq())