UDST / urbansim

94%

master: 94%

LAST BUILD ON BRANCH weird-performance-bug-in-predict

branch: weird-performance-bug-in-predict

CHANGE BRANCH

x

Reset

• weird-performance-bug-in-predict
• appveyor
• dataframe-explorer-port
• deck-parking-developer-issue
• dev
• fix-pandana-0.3.0
• fix-sphinx-issues
• general-profit-to-prob
• installation-cleanup
• lcm_edit
• master
• no-choosers
• no-tall-skinny-buildings
• pandana-0.4.0
• pandana-update
• python3
• release
• release-prep
• return-index-for-dev-merge
• revert-alt-sampling
• sort_bug
• statsmodels
• subsidized-buildings-issue
• tlz
• update_pandas
• v3.0.0
• v3.1.1
• v3.2
• yaml-deprecations

Build # 1111

Build Type

push

travis-ci

Committed by fscottfoti

Commit Message

merging in the UAL changes uncovered this performance issue in UrbanSim

I can't actually explain why, but this one tiny change speeds the simulation up 25 minutes for a 70 minutes simulation.

There should be no behavior difference here - it just protects the case where there are no alternatives, which would return the same empty list you started with inside this loop.  I don't really understand why it would take so long, and in basic tests outside of UrbanSim I can't recreate it.  It seems like a harmless change though, so merging it to master seems innocuous.

Run Details

1 of 1 new or added line in 1 file covered. (100.0%)

1839 of 1951 relevant lines covered (94.26%)

0.94 hits per line