SDC 0.34.0

Changed

• [Alexey Kozlov] Internal: preparing migration to Numba 0.50.1

SDC 0.33.0

Added

• [Alexander Kalistratov] Add default constructor to void_range::iterator (#875)
• [Alexander Kalistratov] Parallel sort via tbb (#844)
• [Alexander Kalistratov] Update README with tbb information (#872)
• [Alexey Kozlov] Adding extension type for pd.RangeIndex (#820)
• [Alexey Kozlov] Optimize getitem operations by checking for same indexes (#800)
• [Elena Totmenina] Impelemented DataFrame at (#738), loc (#788), iloc (#743)
• [Elena Totmenina] Series.add / Series.lt with fill_value (#655)
• [Maria Rubtsowa] DataFrame reset index (#656)
• [Maria Rubtsowa] Implemented Series.skew() (#813)
• [Maria Rubtsowa] Parallel isin (#603)
• [Maria Rubtsowa] Series str contains (#793)
• [Maria Rubtsowa] Series.pct_change() default parameters support (#792)
• [Sergey Pokhodenko] Initial support for Categorical (#819)

Changed

• [Alexander Kalistratov] Enable Azure pipiline CI for all branches (#883)
• [Andres Guzman-Ballen] Remove pin_compatible; introduces conflicts when building (#864)
• [Denis Smirnov, Angelina Kharchevnikova] Refactored DataFrame structure and all the related methods (#870)
• [Elena Totmenina] Add decrefs in boxing (#836)
• [Elena Totmenina] Groupby limitations (#782)
• [Elena Totmenina] Improve boxing (#832)
• [Maria Rubtsowa] Move section about documentation from README to docs (#818)
• [Sergey Pokhodenko] Update to Numba 0.49.1 (#824, #803, #838, #841)
• [Sergey Pokhodenko] Update to PyArrow 0.17.0 (#839, #831)
• [Sergey Pokhodenko] Run tests in separate processes (#833)
• [Vyacheslav Smirnov] Move conda recipe to sdc/conda-recipe folder (#851)
• [Vyacheslav Smirnov] Specify sdc version from channel for examples testing (#837)

Fixed

• [Alexander Kalistratov] Fix typos in docs/compilation (#787)
• [Alexander Kalistratov] gettimeofday for Windows with python 3.6 (#873)
• [Alexey Kozlov] 40 tests are failed on Linux 3.6 (#882)
• [Alexey Kozlov] Fixing perf issue with getitem by Boolean mask (#790)
• [Elena Totmenina] Fixed documentation for Series (#774) and DataFrame (#779)
• [Sergey Pokhodenko] Fixed README details (#786, #789)

Removed

• [Maria Rubtsowa] Cleaned old code (#812, #799, #795, #794)
• [Sergey Pokhodenko] Disabled code for Parquet (#797, #881)
• [Vyacheslav Smirnov] Remove scipy dependency (#798)

Numba 0.46 support

The latest commit which support Numba 0.46. Next commit use Numba from master branch

Initial pre-release of HPAT

Requires the hpat_req_v0.1 branch of IntelLabs Numba.