CLN (PY2): Remove DataFrame/Series.timetuple

[jorisvandenbossche]

I was looking at the sphinx doc build warnings, and of them was complaining about the DataFrame/Series.timetuple attribute. This dummy timetuple attribute was added in #24056, to fix comparison with datetime.datetime.

Apart from the doc build problem (which can probably be fixed in another way as well), this also adds yet another (and for the rest useless / undocumented) attribute to our main classes.

But from a quick test (at least on linux with python 3.7, with Series) this is not needed to make it work. So removing again to see what travis/azure say.

@jbrockmendel do you remember the context of adding this?

[codecov]

Codecov Report

Merging #26062 into master will decrease coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #26062      +/-   ##
==========================================
- Coverage    91.9%   91.89%   -0.01%     
==========================================
  Files         175      175              
  Lines       52485    52484       -1     
==========================================
- Hits        48234    48230       -4     
- Misses       4251     4254       +3

Flag	Coverage Δ
#multiple	90.45% <ø> (-0.01%)	⬇️
#single	40.74% <ø> (-0.15%)	⬇️

Impacted Files	Coverage Δ
pandas/core/generic.py	93.54% <ø> (-0.01%)	⬇️
pandas/io/gbq.py	75% <0%> (-12.5%)	⬇️
pandas/core/frame.py	96.79% <0%> (-0.12%)	⬇️
pandas/util/testing.py	90.72% <0%> (+0.1%)	⬆️

---

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 a07ed59...accbbfe. Read the comment docs.

[codecov]

Codecov Report

Merging #26062 into master will decrease coverage by <.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master   #26062      +/-   ##
==========================================
- Coverage    91.9%   91.89%   -0.01%     
==========================================
  Files         175      175              
  Lines       52485    52506      +21     
==========================================
+ Hits        48234    48249      +15     
- Misses       4251     4257       +6

Flag	Coverage Δ
#multiple	90.45% <ø> (-0.01%)	⬇️
#single	40.73% <ø> (-0.16%)	⬇️

Impacted Files	Coverage Δ
pandas/core/indexes/datetimes.py	96.4% <ø> (-0.01%)	⬇️
pandas/core/arrays/datetimes.py	97.79% <ø> (-0.01%)	⬇️
pandas/core/generic.py	93.54% <ø> (-0.01%)	⬇️
pandas/io/gbq.py	75% <0%> (-12.5%)	⬇️
pandas/core/sparse/frame.py	95.5% <0%> (-0.2%)	⬇️
pandas/compat/__init__.py	77.5% <0%> (-0.15%)	⬇️
pandas/core/frame.py	96.79% <0%> (-0.12%)	⬇️
pandas/plotting/_core.py	83.77% <0%> (-0.09%)	⬇️
pandas/core/dtypes/base.py	100% <0%> (ø)	⬆️
pandas/core/computation/expr.py	96.7% <0%> (ø)	⬆️
... and 6 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 a07ed59...46d4c42. Read the comment docs.

[jbrockmendel]

do you remember the context of adding this?

I was under the impression that it was needed in order to get datetime.__op__(other) to return NotImplemented. If that impression was incorrect, then this may not be needed.

[jorisvandenbossche]

In [1]: import datetime                                                                                                                                       

In [2]: dt = datetime.datetime(2012, 1, 1)                                                                                                                    

In [3]: s = pd.Series([dt, dt + datetime.timedelta(1)])                                                                                                       

In [4]: dt.__eq__(s)                                                                                                                                          
Out[4]: NotImplemented

On this branch (so without the attribute), but in a single env of course.

But I assume that you added tests for it in the original PR, so if the tests are green, we can rely on that?

[jreback]

btw i think was needed for PY2 as that attribute was looked for

[jorisvandenbossche]

Ah, that might be a good reason that CI is green now

[jorisvandenbossche]

Yes, can confirm that in Python 2 it was needed:

In [1]: import datetime

In [2]: dt = datetime.datetime(2012, 1, 1)

In [3]: s = pd.Series([dt, dt + datetime.timedelta(1)])

In [4]: dt == s
Out[4]: False

In [5]: pd.Series.timetuple=None

In [6]: s = pd.Series([dt, dt + datetime.timedelta(1)])

In [7]: dt == s
Out[7]: 
0     True
1    False
dtype: bool

Then we can remove it now. I think there are some other occurrences as well (on the arrays), let me check for that.

[jreback]

thanks @jorisvandenbossche