Description
Code Sample, a copy-pastable example if possible
In [1]: import pandas as pd
In [2]: from pandas.tests.extension.decimal.array import DecimalArray
In [3]: from pandas.tests.extension.json.array import JSONArray
In [4]: DecimalArray([]).isna()
Out[4]: array([], dtype=float64)
In [5]: JSONArray([]).isna()
Out[5]: array([], dtype=float64)
In [6]: import decimal
In [7]: DecimalArray([decimal.Decimal(1.0)]).isna()
Out[7]: array([False])
In [8]: DecimalArray([decimal.Decimal(1.0)]).isna().dtype
Out[8]: dtype('bool')Problem description
This relates to a discussion with @TomAugspurger in #21183 when I made (since redacted) changes to pandas/util/testing.py in assert_extension_array_equal . It turns out that the implementations of isna() in DecimalArray and JSONArray return an ndarray of the wrong dtype if the arrays are empty.
So we need to force the dtype of the result of isna() to be bool in those implementations
Expected Output
In [4]: DecimalArray([]).isna()
Out[4]: array([], dtype=bool)
In [5]: JSONArray([]).isna()
Out[5]: array([], dtype=bool)
Output of pd.show_versions()
INSTALLED VERSIONS
commit: None
python: 3.6.4.final.0
python-bits: 64
OS: Windows
OS-release: 10
machine: AMD64
processor: Intel64 Family 6 Model 60 Stepping 3, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None
LOCALE: None.None
pandas: 0.23.0
pytest: 3.3.2
pip: 9.0.1
setuptools: 38.4.0
Cython: 0.27.3
numpy: 1.14.0
scipy: 1.0.0
pyarrow: None
xarray: None
IPython: 6.2.1
sphinx: 1.6.6
patsy: 0.5.0
dateutil: 2.6.1
pytz: 2017.3
blosc: None
bottleneck: 1.2.1
tables: 3.4.2
numexpr: 2.6.4
feather: None
matplotlib: 2.1.2
openpyxl: 2.4.10
xlrd: 1.1.0
xlwt: 1.3.0
xlsxwriter: 1.0.2
lxml: 4.1.1
bs4: 4.6.0
html5lib: 1.0.1
sqlalchemy: 1.2.1
pymysql: 0.7.11.None
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None