Details
-
Bug
-
Resolution: Unresolved
-
Low
-
None
-
7.4.8
-
None
-
1
-
Severity 3 - Minor
-
Description
Issue Summary
Confluence search ranking doesn't match with that of explain.action .It looks like there is further filtering on results returned by Lucene filter
Steps to Reproduce
- Create a page titled Song
- Create a page with any title and add song in page content
- Create a new page with any title and attachment as Song
Expected Results
Results of explain.action should be similar to that present in Confluence search
Actual Results
In explain.action we can see following,the page titled Song has a higher ranking
query: (title:song^2.0 contentBody:song content-name-unstemmed:song) (1) @@@hl@@@Song@@@endhl@@@: 22.103605 = boostByModificationDate 14.735737 = (MATCH) sum of: 9.309661 = (MATCH) weight(title:song^2.0 in 201) [BM25LSimilarity], result of: 9.309661 = score(doc=201,freq=1.0 = termFreq=1.0), product of: 2.0 = boost 3.63279 = idf(docFreq=5, maxDocs=207) 1.2813376 = tfNorm, computed from: 1.0 = termFreq=1.0 1.25 = parameter k1 0.3 = parameter b 1.7971015 = avgFieldLength 1.0 = fieldLength 5.4260764 = (MATCH) weight(content-name-unstemmed:song in 201) [BM25LSimilarity], result of: 5.4260764 = score(doc=201,freq=1.0 = termFreq=1.0), product of: 4.4212475 = idf(docFreq=2, maxDocs=207) 1.2272727 = tfNorm, computed from: 1.0 = termFreq=1.0 1.25 = parameter k1 0.3 = parameter b 1.0 = avgFieldLength 1.0 = fieldLength (2) @@@hl@@@song@@@endhl@@@: 22.103605 = boostByModificationDate 14.735737 = (MATCH) sum of: 9.309661 = (MATCH) weight(title:song^2.0 in 1) [BM25LSimilarity], result of: 9.309661 = score(doc=1,freq=1.0 = termFreq=1.0), product of: 2.0 = boost 3.63279 = idf(docFreq=5, maxDocs=207) 1.2813376 = tfNorm, computed from: 1.0 = termFreq=1.0 1.25 = parameter k1 0.3 = parameter b 1.7971015 = avgFieldLength 1.0 = fieldLength 5.4260764 = (MATCH) weight(content-name-unstemmed:song in 1) [BM25LSimilarity], result of: 5.4260764 = score(doc=1,freq=1.0 = termFreq=1.0), product of: 4.4212475 = idf(docFreq=2, maxDocs=207) 1.2272727 = tfNorm, computed from: 1.0 = termFreq=1.0 1.25 = parameter k1 0.3 = parameter b 1.0 = avgFieldLength 1.0 = fieldLength (3) ddow: 9.653341 = boostByModificationDate 6.4355607 = (MATCH) sum of: 6.4355607 = (MATCH) weight(contentBody:song in 197) [BM25LSimilarity], result of: 6.4355607 = score(doc=197,freq=1.0 = termFreq=1.0), product of: 4.932073 = idf(docFreq=1, maxDocs=207) 1.3048389 = tfNorm, computed from: 1.0 = termFreq=1.0 1.25 = parameter k1 0.3 = parameter b 6.6280193 = avgFieldLength 2.56 = fieldLength
2) While searching on Confluence UI shows
3) Searching through Luke directly on index is similar to explain.action
Workaround
None