Chun Chang created DRILL-2348: --------------------------------- Summary: 'null' is not treated correctly when compared with int Key: DRILL-2348 URL: https://issues.apache.org/jira/browse/DRILL-2348 Project: Apache Drill Issue Type: Bug Reporter: Chun Chang Priority: Critical
#Wed Feb 25 17:07:31 EST 2015 git.commit.id.abbrev=f7ef5ec Dataset can be downloaded from https://s3.amazonaws.com/apache-drill/files/complex.json.gz The following three query results do not add up. {code} 0: jdbc:drill:schema=dfs.drillTestDirComplexJ> select count(tt.gbyi) from (select t.gbyi gbyi, t.ooa[0] ooa0, t.ooa[1] ooa1, t.ooa[2] ooa2 from `complex.json` t) tt where tt.ooa0.`in` <> tt.ooa1.`in`; +------------+ | EXPR$0 | +------------+ +------------+ No rows selected (22.952 seconds) 0: jdbc:drill:schema=dfs.drillTestDirComplexJ> select count(tt.gbyi) from (select t.gbyi gbyi, t.ooa[0] ooa0, t.ooa[1] ooa1, t.ooa[2] ooa2 from `complex.json` t) tt where tt.ooa0.`in` = tt.ooa1.`in`; +------------+ | EXPR$0 | +------------+ | 949954 | +------------+ 1 row selected (23.053 seconds) 0: jdbc:drill:schema=dfs.drillTestDirComplexJ> select count(tt.gbyi) from (select t.gbyi gbyi, t.ooa[0] ooa0, t.ooa[1] ooa1, t.ooa[2] ooa2 from `complex.json` t) tt; +------------+ | EXPR$0 | +------------+ | 1000000 | +------------+ 1 row selected (13.242 seconds) {code} Without any comparison condition, the total count is 1,000,000. This is correct. But the two query results with <> and = does not add up to the total. I am not sure if this has anything to do with subquery with complex type. Will investigate more. -- This message was sent by Atlassian JIRA (v6.3.4#6332)