Fix output of print() matches with pandas of Series

[beobest2]

Resolve #1248

Since the Python interpreter calls __repr__ if __str__ is not implemented,
the output will be identical to pandas using the implemented __repr__ of series.py.

>>> ks = ks.Series([1, 3, 5, np.nan, 6, 8])
>>> print(ks)
0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
Name: 0, dtype: float64
>>> print(str(ks))
0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
Name: 0, dtype: float64
>>> print(repr(ks))
0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
Name: 0, dtype: float64
>>> ps = pd.Series([1, 3, 5, np.nan, 6, 8])
>>> print(ps)
0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
dtype: float64
>>> print(str(ps))
0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
dtype: float64
>>> print(repr(ps))
0    1.0
1    3.0
2    5.0
3    NaN
4    6.0
5    8.0
dtype: float64

[codecov-io]

Codecov Report

Merging #1250 into master will increase coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff            @@
##           master   #1250      +/-   ##
=========================================
+ Coverage   95.09%   95.1%   +0.01%     
=========================================
  Files          35      35              
  Lines        7154    7152       -2     
=========================================
- Hits         6803    6802       -1     
+ Misses        351     350       -1

Impacted Files	Coverage Δ
databricks/koalas/series.py	96.37% <ø> (+0.13%)	⬆️
databricks/koalas/frame.py	96.83% <0%> (-0.01%)	⬇️
databricks/koalas/ml.py	96.29% <0%> (+0.14%)	⬆️

---

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 dfd3bd1...ba559bb. Read the comment docs.

[ueshin]

LGTM.
Actually I don't remember why we use the string.
@HyukjinKwon do you know the reason?

[itholic]

LGTM, too if there is no proper reason we should use the existing behavior.

[itholic]

nit: we better avoid ks as a variable name of Series since there is a high probability of overwriting the Koalas package name, ks. (I recommend kser instead, generally)

[beobest2]

nit: we better avoid ks as a variable name of Series since there is a high probability of overwriting the Koalas package name, ks. (I recommend kser instead, generally)

Thanks for the good tip :)

[HyukjinKwon]

Looks fine given that it's consistent with DataFrame. I don't know if there's the particular reason either.

One concern might be it triggers an actual job when, for instance, the string representation is expected in an error message .. but I suspect it's fine for now.