Was reviewing the accepted answer of another question.
The section in the demonstration that confuses me is
In [37]: (sign != sign.shift()).cumsum()
Out[37]:
0 1
1 1
2 2
3 2
4 2
5 3
6 3
Name: values, dtype: int64
It appears that everytime the (sign != sign.shift()) gives a True result, cumsum() returns an incremented number, and keeps returning that number until it encounters another True.
Looking at the sparse docs for .cumsum, I have no idea how/why that behavior is occuring, however useful that it is!
My only guess is cumsum() treats True as 1 and False as 0. Doesn't seem like a pythonic thing to do -- sounds like something from C.
Could someone explain why this effect occurs?
