[WIP]: Tweak Int64Index.reindex() performance

IntelPython / sdc

Numba extension for compiling Pandas data frames, Intel® Scalable Dataframe Compiler

https://intelpython.github.io/sdc-doc/

BSD 2-Clause "Simplified" License

645 stars 61 forks source link

[WIP]: Tweak Int64Index.reindex() performance #984

Open kozlov-alexey opened 2 years ago

kozlov-alexey commented 2 years ago

Motivation: current implementation of Int64Index reindex via building map_positions as a typed.Dict and filling the result indexer in a prange scales poorly and has performance that is far from ideal. This PR improves performance by ~2x by using native TBB based implementation.

pep8speaks commented 2 years ago

Hello @kozlov-alexey! Thanks for opening this PR. We checked the lines you've touched for PEP 8 issues, and found:

In the file sdc/extensions/sdc_hashmap_ext.py:

Line 69:1: E266 too many leading '#' for block comment Line 71:1: E302 expected 2 blank lines, found 0 Line 76:1: E305 expected 2 blank lines after class or function definition, found 1 Line 1175:59: E231 missing whitespace after ','