[ https://issues.apache.org/jira/browse/OAK-3230?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14712917#comment-14712917 ]
Thomas Mueller commented on OAK-3230: ------------------------------------- Ah oak-lucene fails with a NPE, so we do have enough tests: {noformat} testNtFile(org.apache.jackrabbit.oak.jcr.query.TextExtractionQueryTest) Time elapsed: 0.207 sec <<< ERROR! java.lang.NullPointerException at org.apache.jackrabbit.oak.plugins.index.aggregate.AggregationCursor.fetchNext(AggregationCursor.java:91) at org.apache.jackrabbit.oak.plugins.index.aggregate.AggregationCursor.hasNext(AggregationCursor.java:76) at org.apache.jackrabbit.oak.query.ast.SelectorImpl.next(SelectorImpl.java:402) at org.apache.jackrabbit.oak.query.QueryImpl$RowIterator.fetchNext(QueryImpl.java:780) at org.apache.jackrabbit.oak.query.QueryImpl$RowIterator.hasNext(QueryImpl.java:805) at org.apache.jackrabbit.oak.jcr.query.QueryResultImpl$4.fetch(QueryResultImpl.java:172) {noformat} > Yes, virtual row check should have been fine against nextRow. I didn't want > to update currentRow directly with cursor.next() as it seemed current row for > aggregation cursor shouldn't be virtual. Yes aggregation in combination with virtual rows is unexpected. But I guess queries that produce virtual rows don't use aggregation, so we should be fine. > Query engine should support virtual index rows > ---------------------------------------------- > > Key: OAK-3230 > URL: https://issues.apache.org/jira/browse/OAK-3230 > Project: Jackrabbit Oak > Issue Type: Sub-task > Components: query > Reporter: Vikas Saurabh > Assignee: Thomas Mueller > Fix For: 1.3.5 > > Attachments: OAK-3230-query-engine-should-support-virtual-rows.patch > > > As discussed in OAK-3156 > [here|https://issues.apache.org/jira/browse/OAK-3156?focusedCommentId=14645712&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14645712] > and > [here|https://issues.apache.org/jira/browse/OAK-3156?focusedCommentId=14655273&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14655273], > we need support for virtual rows returned by indices (rows where path > property doesn't make sense). > Talked to [~tmueller] and we should have a new method in {{IndexRow}} to > clearly mark the intent that the row is virtual and hence should be treated > accordingly by the query engine. -- This message was sent by Atlassian JIRA (v6.3.4#6332)