Inconsistency, NaT included in result of groupby method first but not NaN

[larvian]

NaT is included in result of groupby method first while NaN. I am expecting that first should skip both NaN and NaT and include the first value where pandas.isnull is False.
Demonstration of the inconsistency. (note that both NaT and NaN in the data frame are produced by np.nan, the difference is that the d_t column contains date values).

import numpy as np
import pandas as pd
from datetime import datetime as dt

testFrame=DataFrame({'IX':['A','A'],'num':[np.nan,100],'d_t':[np.nan,dt.now()]})

Resulting data frame:

  IX                     d_t  num
0  A                     NaT  NaN
1  A 2015-07-15 22:47:10.635  100

Grouping this data frame on the IX column and executing the first method results in this data frame which shows the inconsistency between the d_t and num columns.

testFrame.groupby('IX').first()

Resulting dataframe:

        d_t  num
IX              
A       NaT  100

INSTALLED VERSIONS
------------------
commit: None
python: 2.7.9.final.0
python-bits: 64
OS: Windows
OS-release: 7
machine: AMD64
processor: Intel64 Family 6 Model 58 Stepping 9, GenuineIntel
byteorder: little
LC_ALL: None
LANG: None

pandas: 0.16.2
nose: 1.3.4
Cython: 0.22
numpy: 1.9.2
scipy: 0.15.1
statsmodels: 0.6.1
IPython: 3.0.0
sphinx: 1.2.3
patsy: 0.3.0
dateutil: 2.4.2
pytz: 2015.4
bottleneck: None
tables: 3.1.1
numexpr: 2.3.1
matplotlib: 1.4.3
openpyxl: 1.8.5
xlrd: 0.9.3
xlwt: 0.7.5
xlsxwriter: 0.6.7
lxml: 3.4.2
bs4: 4.3.2
html5lib: 0.999
httplib2: None
apiclient: None
sqlalchemy: 0.9.9
pymysql: None
psycopg2: None

[jreback]

hmm, @sinhrks I thought this was fixed?

[sinhrks]

I remember some issues when NaT is in group key, but not aware the aggregation issure. .min might be affected also.

[jreback]

prob needs some adjustment for comparison vs iNaT in the first/last (though i thought it was there)

[sinhrks]

Confirmed .min also affected. @larvian PR is appreciated:)

testFrame.groupby('IX').min()
#    d_t  num
# IX         
# A  NaT  100

[larvian]

@sinhrks :) I see. Well, I have started to research GitHub and the Pandas source code a bit but right now I unfortunately don't have very much time available and waiting for me could require patience.
If I come to the point when I feel confident that I can contribute to the solution then I will do a PR. If someone else does it I am OK with that too :)