BUG: ARIMA fit with trend and constant exog

[jbrockmendel]

• closes ARIMA ValueError with constant exog variable #3343
• tests added / passed.
• code/documentation is well formatted.
• properly formatted commit message. See
  [NumPy's guide]

ARIMA analogue for #5802.

[bashtage]

Are the failing tests buggy?

[jbrockmendel]

Are the failing tests buggy?

It looks like there are two existing tests that hit this code and now raise. I'll update those tests, since I'm confident they weren't trying to hit this case.

[coveralls]

Coverage decreased (-0.0009%) to 85.033% when pulling 237b653 on jbrockmendel:arima_fit into 744f6cb on statsmodels:master.

[codecov]

Codecov Report

Merging #5833 into master will increase coverage by 0.04%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #5833      +/-   ##
==========================================
+ Coverage   82.45%   82.49%   +0.04%     
==========================================
  Files         597      597              
  Lines       94148    94156       +8     
  Branches     9667    10401     +734     
==========================================
+ Hits        77626    77678      +52     
+ Misses      14095    14058      -37     
+ Partials     2427     2420       -7

Impacted Files	Coverage Δ
statsmodels/tsa/arima_model.py	86.43% <100%> (ø)	⬆️
statsmodels/tsa/tests/test_arima.py	99.51% <100%> (ø)	⬆️
statsmodels/sandbox/distributions/mv_normal.py	39.21% <0%> (ø)	⬆️
statsmodels/sandbox/stats/multicomp.py	40.6% <0%> (ø)	⬆️
statsmodels/sandbox/bspline.py	0% <0%> (ø)	⬆️
statsmodels/genmod/tests/test_glm.py	99.07% <0%> (+0.16%)	⬆️
statsmodels/regression/mixed_linear_model.py	90% <0%> (+0.17%)	⬆️
statsmodels/sandbox/stats/diagnostic.py	65.5% <0%> (+0.21%)	⬆️
statsmodels/graphics/regressionplots.py	81.62% <0%> (+0.23%)	⬆️
... and 10 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 551671a...9110498. Read the comment docs.

[codecov]

Codecov Report

Merging #5833 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff            @@
##           master   #5833      +/-   ##
=========================================
- Coverage    82.5%   82.5%   -0.01%     
=========================================
  Files         597     597              
  Lines       94171   94179       +8     
  Branches    10399   10399              
=========================================
+ Hits        77697   77701       +4     
- Misses      14057   14059       +2     
- Partials     2417    2419       +2

Impacted Files	Coverage Δ
statsmodels/tsa/arima_model.py	86.36% <100%> (-0.5%)	⬇️
statsmodels/tsa/tests/test_arima.py	99.51% <100%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 744f6cb...237b653. Read the comment docs.

[bashtage]

Can you rebase? I think this is ready.

[bashtage]

I changed the error message to make the problem more obvious, and I think your rebase changed the test back., hence the fail.

[jbrockmendel]

Rebase-induced test failures fixed

[bashtage]

Thanks.