reindex() inconsistency between Index and MultiIndex

[matthewgilbert]

This seems related but distinct from #7895

import pandas as pd
from pandas import Timestamp as TS

df = pd.DataFrame([1, 2, 3], index=[0, 2, 4])
 # throws an error as expected
df.reindex([0, 1, 2, 3, 4, 5, 6], fill_value=0, limit=1)

ValueError: limit argument only valid if doing pad, backfill or nearest reindexing

idx = pd.MultiIndex.from_tuples([(TS('2015-01-03'), 'A'),
                                 (TS('2015-01-04'), 'A'),
                                 (TS('2015-01-04'), 'B')])
df2 = pd.Series([0.02, 0.01, 0.012], index=idx)

widx = pd.MultiIndex.from_tuples([(TS('2015-01-03'), 'A'),
                                  (TS('2015-01-03'), 'B'),
                                  (TS('2015-01-04'), 'A'),
                                  (TS('2015-01-04'), 'B'),
                                  (TS('2015-01-04'), 'C'),
                                  (TS('2015-01-05'), 'B'),
                                  (TS('2015-01-05'), 'C'),
                                  (TS('2015-01-06'), 'B'),
                                  (TS('2015-01-06'), 'C')])
df2.reindex(widx, fill_value=0, limit=1)

2015-01-03  A    0.020
            B    0.000
2015-01-04  A    0.010
            B    0.012
            C    0.000
2015-01-05  B    0.000
            C    0.000
2015-01-06  B    0.000
            C    0.000
dtype: float64

Problem description

The behaviour illustrated above of reindex on a MultiIndex appears to be a bug to me. I would not expect reindex to support limit with a fill_value for a MultiIndex but not an Index, and in addition it appears like this limit is accepted but ignored for a MultiIndex reindex.

Output of pd.show_versions()

INSTALLED VERSIONS

commit: None
python: 3.6.0.final.0
python-bits: 64
OS: Linux
OS-release: 4.15.0-22-generic
machine: x86_64
processor: x86_64
byteorder: little
LC_ALL: None
LANG: en_CA.UTF-8
LOCALE: en_CA.UTF-8

pandas: 0.23.0
pytest: 3.0.5
pip: 9.0.1
setuptools: 27.2.0
Cython: 0.25.2
numpy: 1.13.1
scipy: 0.19.1
pyarrow: 0.8.0
xarray: None
IPython: 5.1.0
sphinx: 1.5.4
patsy: 0.4.1
dateutil: 2.6.0
pytz: 2016.10
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.2
feather: None
matplotlib: 2.0.2
openpyxl: 2.4.1
xlrd: 1.0.0
xlwt: 1.2.0
xlsxwriter: 0.9.6
lxml: 3.7.2
bs4: 4.5.3
html5lib: 0.9999999
sqlalchemy: 1.1.5
pymysql: None
psycopg2: None
jinja2: 2.9.4
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: 0.2.1

[mroeschke]

This looks to raise as expected on master now. Could use a test

In [4]: idx = pd.MultiIndex.from_tuples([(TS('2015-01-03'), 'A'),
   ...:                                  (TS('2015-01-04'), 'A'),
   ...:                                  (TS('2015-01-04'), 'B')])
   ...: df2 = pd.Series([0.02, 0.01, 0.012], index=idx)
   ...:
   ...: widx = pd.MultiIndex.from_tuples([(TS('2015-01-03'), 'A'),
   ...:                                   (TS('2015-01-03'), 'B'),
   ...:                                   (TS('2015-01-04'), 'A'),
   ...:                                   (TS('2015-01-04'), 'B'),
   ...:                                   (TS('2015-01-04'), 'C'),
   ...:                                   (TS('2015-01-05'), 'B'),
   ...:                                   (TS('2015-01-05'), 'C'),
   ...:                                   (TS('2015-01-06'), 'B'),
   ...:                                   (TS('2015-01-06'), 'C')])
   ...: df2.reindex(widx, fill_value=0, limit=1)
ValueError: limit argument only valid if doing pad, backfill or nearest reindexing

[calvh]

take