Github user ilaygit commented on a diff in the pull request: https://github.com/apache/lucene-solr/pull/162#discussion_r185438730 --- Diff: solr/core/src/java/org/apache/solr/handler/component/QueryComponent.java --- @@ -1414,10 +1267,17 @@ private void doProcessGroupedDistributedSearchFirstPhase(ResponseBuilder rb, Que .setSearcher(searcher); for (String field : groupingSpec.getFields()) { + final int topNGroups; + Query query = cmd.getQuery(); + if (query instanceof AbstractReRankQuery){ + topNGroups = cmd.getOffset() + ((AbstractReRankQuery)query).getReRankDocs(); + } else { + topNGroups = cmd.getOffset() + cmd.getLen(); + } topsGroupsActionBuilder.addCommandField(new SearchGroupsFieldCommand.Builder() .setField(schema.getField(field)) .setGroupSort(groupingSpec.getGroupSort()) - .setTopNGroups(cmd.getOffset() + cmd.getLen()) --- End diff -- Is it reRanking with respect to the "start" param in the solr query? Looks like the original LTR plugin is supposed to reRank top K irrespective of the offset. ref.http://lucene.472066.n3.nabble.com/Learning-to-Rank-LTR-with-grouping-td4366691i20.html#a4387929
--- --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org