Pandafy data

[florian-huber]

alternative route to same goal. Now using panda multiindex DataFrame underneath.

[florian-huber]

Up to this point:
The panda-based solution works as well.
Pros:

• Code is more readable
• Stacking arrays is much more flexible ("left", "right", "inner", "outer" joins are possible)
• Stacking arrays is much faster.

Cons:

• Slicing became much slower! --> Maybe adapt slicing more to pandas implementation?
  (one option might be to use df._values directly instead of going via .iloc)

[florian-huber]

After a lot of profiling I decided against a Pandas implementation (too slow slicing) --> See #11 .