[ https://issues.apache.org/jira/browse/KYLIN-5782?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17834933#comment-17834933 ]
pengfei.zhan commented on KYLIN-5782: ------------------------------------- h1. Root Cause After the context is sliced, the columns on the join condition will be pressed down to the sub context, the customer's join condition is like a = b and c = d and a = e, the code is using map to collect the information of the join keys, and use the column on the left of the equals sign as the key, and the column on the right as the value, when a=e is put into the map, the column b is overwritten due to the duplication of the key and a=b, and only the column c, d, a, and e are collected, and the column b is omitted. When a=e is put into the map, b is overwritten because the key is duplicated with a=b, and the only columns collected are c, d, a, and e, leaving out b. This results in b being used as the dummy column, but this column will eventually be used as the join key, which results in the data being empty. > Duplicated join keys result empty query result > ---------------------------------------------- > > Key: KYLIN-5782 > URL: https://issues.apache.org/jira/browse/KYLIN-5782 > Project: Kylin > Issue Type: Bug > Components: Query Engine > Affects Versions: 5.0-beta > Reporter: pengfei.zhan > Assignee: pengfei.zhan > Priority: Major > Fix For: 5.0-beta > > -- This message was sent by Atlassian Jira (v8.20.10#820010)