eriknw / toolz

100%

master: 100%

LAST BUILD ON BRANCH faster_groupby

branch: faster_groupby

CHANGE BRANCH

x

Reset

• faster_groupby
• authors_larsmans_josericardo
• authors_shvechikov
• backport_tests
• better_curry_error
• better_curry_introspection
• clean_curried_namespace
• consume
• correct_flip_doctest
• coveralls
• coveralls_for_master
• coveralls_success_only
• curry_class
• curry_classmethods
• curry_makeover
• curry_memoize
• curry_numargs
• curry_reduce
• dict_iter
• diff
• equality_hash_key
• explicit_curried
• faster_merge_sorted
• faster_merge_with
• faster_remove
• faster_unique
• fix/curry_module
• fix/ensure_easy_testing
• fix_32bit_overflow
• fix_property_introspection
• get_with_itemgetter
• identical
• isdistinct_for_iterators
• itemmap
• jackknife_sandbox
• juxt_generator
• juxt_tuple
• last_with_deque
• memoize_key_func
• memoize_unary
• more_introspection
• nth_with_list
• pep479
• pypy3
• python3.6
• readme_broken_links
• readme_rst
• remove_jackknife
• run_doctests_on_curried
• serialize_juxt
• svg_badges
• test_curry_too
• test_dicttoolz_types
• tlz
• unused_variables
• update/pluck_docstring

Build # 92

Build Type

push

travis-ci

Committed by eriknw

Commit Message

Choose a version of `groupby` that is easier to understand.

This is `groupby_alt` from the original commit comment in this branch.
See discussion at #179.  This version performs very well as groups become
larger.  The previous implementation performs well when 10% or more of
the elements form groups of length one.

We plan to have various implementations in `benchmarkz` repository, which
will let contributors add benchmarks that they care about and easily run
them on all variants of `groupby`.

Run Details

458 of 458 relevant lines covered (100.0%)

4.0 hits per line