ENH: Implement '_period_dtype_code' attribute to 'pandas._libs.tslibs.offsets.MonthBegin'

[yohplala]

Code...

index_offsets = pd.tseries.frequencies.to_offset('1MS')
time_marker = pd.Timestamp('2020/01/02 1:59')
period = pd.Period(time_marker, freq = index_offsets.freqstr)

... raises following error message:

Traceback (most recent call last):

  File "pandas/_libs/tslibs/period.pyx", line 2436, in pandas._libs.tslibs.period.freq_to_dtype_code

AttributeError: 'pandas._libs.tslibs.offsets.MonthBegin' object has no attribute '_period_dtype_code'


The above exception was the direct cause of the following exception:

Traceback (most recent call last):

  File "<ipython-input-63-c78e6c74b47c>", line 1, in <module>
    period = pd.Period(time_marker, freq = index_offsets.freqstr)

  File "pandas/_libs/tslibs/period.pyx", line 2422, in pandas._libs.tslibs.period.Period.__new__

  File "pandas/_libs/tslibs/period.pyx", line 2438, in pandas._libs.tslibs.period.freq_to_dtype_code

ValueError: Invalid frequency: {0}

Please, could an appropriate '_period_dtype_code' attribute be provided to 'pandas._libs.tslibs.offsets.MonthBegin' object to solve this error?
Thanks for your amazing lib!

INSTALLED VERSIONS
------------------
commit           : db08276bc116c438d3fdee492026f8223584c477
python           : 3.8.5.final.0
python-bits      : 64
OS               : Linux
OS-release       : 5.4.0-58-generic
Version          : #64-Ubuntu SMP Wed Dec 9 08:16:25 UTC 2020
machine          : x86_64
processor        : x86_64
byteorder        : little
LC_ALL           : None
LANG             : fr_FR.UTF-8
LOCALE           : fr_FR.UTF-8

pandas           : 1.1.3
numpy            : 1.19.2
pytz             : 2020.1
dateutil         : 2.8.1
pip              : 20.2.4
setuptools       : 50.3.1.post20201107
Cython           : 0.29.21
pytest           : 6.1.1
hypothesis       : None
sphinx           : 3.2.1
blosc            : None
feather          : None
xlsxwriter       : 1.3.7
lxml.etree       : 4.6.1
html5lib         : 1.1
pymysql          : None
psycopg2         : None
jinja2           : 2.11.2
IPython          : 7.19.0
pandas_datareader: None
bs4              : 4.9.3
bottleneck       : 1.3.2
fsspec           : 0.8.3
fastparquet      : 0.4.2
gcsfs            : None
matplotlib       : 3.3.2
numexpr          : 2.7.1
odfpy            : None
openpyxl         : 3.0.5
pandas_gbq       : None
pyarrow          : None
pytables         : None
pyxlsb           : None
s3fs             : None
scipy            : 1.5.2
sqlalchemy       : 1.3.20
tables           : 3.6.1
tabulate         : None
xarray           : None
xlrd             : 1.2.0
xlwt             : 1.3.0
numba            : 0.51.2

Not building from sources (using anaconda)

[jreback]

you need to show versions and if you are building from source

[yohplala]

you need to show versions and if you are building from source

Hi Jeff, sorry, information added.

[jbrockmendel]

There is no Period corresponding to MonthBegin. Monthly Periods use offsets.MonthEnd

[yohplala]

There is no Period corresponding to MonthBegin. Monthly Periods use offsets.MonthEnd

Hi, I have tried to check in the source, and I am not finding any MonthBegin in pandas/_libs/tslibs/period.pyxas stated by the error message, so my answer is maybe not appropriate.

Your answer seems to relate to the 2nd part of the error message:
File "pandas/_libs/tslibs/period.pyx", line 2438, in pandas._libs.tslibs.period.freq_to_dtype_code

But what happens if the direct root cause is solved?
AttributeError: 'pandas._libs.tslibs.offsets.MonthBegin' object has no attribute '_period_dtype_code'

Why not adding a Month '_period_dtype_code` to 'pandas._libs.tslibs.offsets.MonthBegin' object?
What does prevent it?

[jbrockmendel]

There is not a one-to-one mapping between DateOffset subclasses and Periods. There are many DateOffsets for which there is no corresponding Period. MonthBegin is one of them.

[yohplala]

There is not a one-to-one mapping between DateOffset subclasses and Periods. There are many DateOffsets for which there is no corresponding Period. MonthBegin is one of them.

Yes I understand that, and I can understand that a Period can only point toward single DateOffset.
But my question is
'What is the a trouble if several DateOffsets point toward the same Period?'

[jbrockmendel]

@MarcoGorelli has the recent work in documenting what offsets are valid Periods address this?

[MarcoGorelli]

Almost

All that needs doing to close this issue IMO is to, in

pandas/pandas/_libs/tslibs/period.pyx

Lines 2675 to 2677 in 39c69da

	One of pandas period strings or corresponding objects. Accepted
	strings are listed in the
	:ref:`offset alias section <timeseries.offset_aliases>` in the user docs.

to replace "offset alias" with "period alias", and to change the link to timeseries.period_aliases

Anyone fancy making a PR?

[nandsha]

Hi @MarcoGorelli ,

Thanks for the opportunity!
I would like to (start) contribute, and after going through the details, I find this one seems to be a good fit for me!
I will try to get back on this with an update in about 2-3 weeks' time, and if that sounds fair, could you please assign this one to me?

Cheers,
Nandha

[MarcoGorelli]

all that's needed is a 2-word change, it shouldn't take 2-3 weeks' time

if anyone wants to make a PR, feel free to do so

[rsm-23]

Can I take it @MarcoGorelli ? Or should I wait for @nandsha to comeback?

[MarcoGorelli]

go ahead, thanks @rsm-23

[rsm-23]

take