[LANG-1199] Fix implementation of StringUtils.getJaroWinklerDistance()

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.4
• Fix Version/s: 3.5
• Component/s: lang.*
• Assignee: Pascal Schumacher
• Reporter: M. Steiger
• Created: 13/Jan/16 15:54
• Updated: 03/Apr/17 12:45
• Resolved: 05/Jun/16 15:40

Description

The current implementation of StringUtils.getJaroWinklerDistance() does not compute the correct result in some cases. See #LANG-944 for the initial code contribution.

StringUtils.getJaroWinklerDistance("Haus Ingeborg", "Ingeborg Esser") == 0.0

This is due to the incorrect computation of common characters, which causes the algorithm to exit prematurely.

In contrast, the implementation in Lucene gives ~0.63, which is about right.

JaroWinklerDistance d = new JaroWinklerDistance();
getDistance("Haus Ingeborg", "Ingeborg Esser");

See https://lucene.apache.org/core/3_0_3/api/contrib-spellchecker/org/apache/lucene/search/spell/JaroWinklerDistance.html

Issue Links

• is related to: TEXT-76 Jaro Winkler implementation is not correct
• relates to: TEXT-26 Check if LANG-1199 also affects commons-text
• links to: GitHub Pull Request #148