Added ``via_values`` property to ``Series``, ``Index``, ``Frame``, ``IndexHierarchy`` and ``Batch``; permits applying functions to complete containers with ``apply()`` and supports usage as arguments in arbitrary NumPy functions with ``__array_ufunc__()``.
Corrected usage of ``IterNodeDelegate`` with iterator endpoints that do not iterate hashables; added ``IterNodeDelegateMapable`` for usage with iterators of hashables.
Improved type and dtype preservation in concatenation and set operations on ``IndexHierarchy``.
Normalized ordering of results from ``Frame.bloc[]`` selections to row-major ordering without sorting labels.
Added ``via_str.contains()``.
Corrected issue in ``ArchiveZIP`` when ``__del__`` is called when no archive is set.
``Frame.to_sqlite()`` now requires a named ``Frame`` or an explicit ``label``; ``Frame.from_sqlite()``, ``Frame.from_hdf5()`` now make ``label`` a required argument.
API Documentation re-organized, now using procedurally generated code examples.