Github user dsmiley commented on a diff in the pull request: https://github.com/apache/lucene-solr/pull/385#discussion_r191978869 --- Diff: solr/core/src/java/org/apache/solr/update/DirectUpdateHandler2.java --- @@ -417,7 +417,8 @@ private void addAndDelete(AddUpdateCommand cmd, List<UpdateLog.DBQ> deletesAfter } private Term getIdTerm(AddUpdateCommand cmd) { --- End diff -- Latest diff shows this logic calls cmd.computeFlattenedDocs() and we definitely don't want this method computing that! Now that I look at the code in my IDE, I can appreciate that this is a bit of a tricky issue though, since neither callers of getIdTerm yet have the List<SolrInputDocument> to give to this method. Hmmmm. Maybe `updateDocument` should not take an updateTerm as an argument but should instead figure it out and return it? Note that if isInPlaceUpdate, the updateTerm will always be the unique key field; you needn't check for a list of docs.
--- --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org