Series[sparse].to_sparse messes up length

```python
In [23]: a = pd.Series(pd.SparseArray([1, 2, None, 2]))

In [24]: b = a.to_sparse()

In [25]: len(a), len(b)
Out[25]: (4, 3)

In [26]: b
Out[26]:
0    1.0
1    2.0
2    2.0
3
dtype: float64
IntIndex
Indices: array([0, 1, 2], dtype=int32)
```

Fixing in https://github.com/pandas-dev/pandas/pull/22325, this is to link from the release notes.