Concat fails with tz-aware and empty series (axis=1)

[linar-jether]

Code Sample, a copy-pastable example if possible

# This works as expected, and as it worked in the past (0.17.1)
pd.concat([pd.util.testing.makeTimeSeries(), pd.Series()], axis=1)

# This fails in 0.20.1
pd.concat([pd.util.testing.makeTimeSeries().tz_localize('America/Chicago'), pd.Series()], axis=1)

Problem description

Only happens with axis=1
Shouldn't be a problem to concat an empty series to a dataframe...

Expected Output

Should filter out empty objects passed to concat.

pd.concat([pd.Series(pd.util.testing.makeTimeSeries().tz_localize('America/Chicago'), name='with_tz'), pd.Series(name='empty')], axis=1)
Out[2]: 
                            with_tz
2000-01-03 00:00:00-06:00 -0.999300
2000-01-04 00:00:00-06:00 -0.800759
2000-01-05 00:00:00-06:00  0.166496
2000-01-06 00:00:00-06:00 -0.496464
2000-01-07 00:00:00-06:00 -0.237731
2000-01-10 00:00:00-06:00 -0.242352
2000-01-11 00:00:00-06:00  1.779163
2000-01-12 00:00:00-06:00  0.193893

Output of pd.show_versions()

``` INSTALLED VERSIONS ------------------ commit: None python: 2.7.12.final.0 python-bits: 64 OS: Windows OS-release: 8.1 machine: AMD64 processor: Intel64 Family 6 Model 60 Stepping 3, GenuineIntel byteorder: little LC_ALL: None LANG: None LOCALE: None.None pandas: 0.20.2 pytest: 3.0.7 pip: 9.0.1 setuptools: 34.3.1 Cython: None numpy: 1.12.1 scipy: 0.18.1 xarray: None IPython: 5.3.0 sphinx: None patsy: 0.4.1 dateutil: 2.6.0 pytz: 2017.2 blosc: None bottleneck: None tables: None numexpr: 2.6.2 feather: None matplotlib: 2.0.2 openpyxl: None xlrd: 1.0.0 xlwt: None xlsxwriter: None lxml: None bs4: 4.5.1 html5lib: 0.9999999 sqlalchemy: None pymysql: None psycopg2: None jinja2: 2.9.6 s3fs: None pandas_gbq: None pandas_datareader: None

[chris-b1]

On your expected output - shouldn't there be an 'empty' column filled with NaNs, like the non tz case?

For consistency, I agree this should work, although in practice it's sort of odd behavior to rely on, I'd be more inclined to explicitly filter, like this. Still, thanks for the report and PRs welcome!

In [15]: pd.concat([x for x in series if not x.empty], axis=1)

[chris-b1]

#16230 might be related

[jreback]

this is a duplicate of #16230. PR's to fix that would be great!