[jira] [Commented] (PHOENIX-3014) SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables
[ https://issues.apache.org/jira/browse/PHOENIX-3014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15343135#comment-15343135 ] Hudson commented on PHOENIX-3014: - FAILURE: Integrated in Phoenix-master #1282 (See [https://builds.apache.org/job/Phoenix-master/1282/]) PHOENIX-3014 SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results (larsh: rev cd3868ce418ded3391be45f89c7c428afd6f520e) * phoenix-core/src/it/java/org/apache/phoenix/end2end/DistinctPrefixFilterIT.java * phoenix-core/src/main/java/org/apache/phoenix/execute/AggregatePlan.java > SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables > -- > > Key: PHOENIX-3014 > URL: https://issues.apache.org/jira/browse/PHOENIX-3014 > Project: Phoenix > Issue Type: Bug >Reporter: Lars Hofhansl >Assignee: James Taylor > Fix For: 4.8.0 > > Attachments: 3014-v2.txt, PHOENIX-3014_untested.patch > > > {code} > create table T(pk1 varchar not null, pk2 varchar not null, constraint pk > primary key(pk1, pk2)) SALT_BUCKETS=8; > upsert into T values('1','1'); > upsert into T values('1','2'); > upsert into T values('2','1'); > select /*+ RANGE_SCAN */ distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > select distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > +--+ > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-3014) SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables
[ https://issues.apache.org/jira/browse/PHOENIX-3014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15342864#comment-15342864 ] Samarth Jain commented on PHOENIX-3014: --- +1 > SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables > -- > > Key: PHOENIX-3014 > URL: https://issues.apache.org/jira/browse/PHOENIX-3014 > Project: Phoenix > Issue Type: Bug >Reporter: Lars Hofhansl > Fix For: 4.8.0 > > Attachments: 3014-v2.txt, PHOENIX-3014_untested.patch > > > {code} > create table T(pk1 varchar not null, pk2 varchar not null, constraint pk > primary key(pk1, pk2)) SALT_BUCKETS=8; > upsert into T values('1','1'); > upsert into T values('1','2'); > upsert into T values('2','1'); > select /*+ RANGE_SCAN */ distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > select distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > +--+ > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-3014) SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables
[ https://issues.apache.org/jira/browse/PHOENIX-3014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15342807#comment-15342807 ] Lars Hofhansl commented on PHOENIX-3014: tests run with {{mvn package}} pass with this and PHOENIX-3012 applied. > SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables > -- > > Key: PHOENIX-3014 > URL: https://issues.apache.org/jira/browse/PHOENIX-3014 > Project: Phoenix > Issue Type: Bug >Reporter: Lars Hofhansl > Fix For: 4.8.0 > > Attachments: 3014-v2.txt, PHOENIX-3014_untested.patch > > > {code} > create table T(pk1 varchar not null, pk2 varchar not null, constraint pk > primary key(pk1, pk2)) SALT_BUCKETS=8; > upsert into T values('1','1'); > upsert into T values('1','2'); > upsert into T values('2','1'); > select /*+ RANGE_SCAN */ distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > select distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > +--+ > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-3014) SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables
[ https://issues.apache.org/jira/browse/PHOENIX-3014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15342768#comment-15342768 ] Lars Hofhansl commented on PHOENIX-3014: Actually removing this part from the patch makes things work: {code} this.getTableRef().getTable().getBucketNum() != null ? SaltingUtil.NUM_SALTING_BYTES : 0, {code} > SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables > -- > > Key: PHOENIX-3014 > URL: https://issues.apache.org/jira/browse/PHOENIX-3014 > Project: Phoenix > Issue Type: Bug >Reporter: Lars Hofhansl > Fix For: 4.8.0 > > Attachments: PHOENIX-3014_untested.patch > > > {code} > create table T(pk1 varchar not null, pk2 varchar not null, constraint pk > primary key(pk1, pk2)) SALT_BUCKETS=8; > upsert into T values('1','1'); > upsert into T values('1','2'); > upsert into T values('2','1'); > select /*+ RANGE_SCAN */ distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > select distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > +--+ > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-3014) SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables
[ https://issues.apache.org/jira/browse/PHOENIX-3014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15342609#comment-15342609 ] Lars Hofhansl commented on PHOENIX-3014: [~samarthjain], this looks like a pretty important issue with incorrect results. > SELECT DISTINCT pk ORDER BY pk DESC gives the wrong results with salted tables > -- > > Key: PHOENIX-3014 > URL: https://issues.apache.org/jira/browse/PHOENIX-3014 > Project: Phoenix > Issue Type: Bug >Reporter: Lars Hofhansl > Fix For: 4.8.0 > > Attachments: PHOENIX-3014_untested.patch > > > {code} > create table T(pk1 varchar not null, pk2 varchar not null, constraint pk > primary key(pk1, pk2)) SALT_BUCKETS=8; > upsert into T values('1','1'); > upsert into T values('1','2'); > upsert into T values('2','1'); > select /*+ RANGE_SCAN */ distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > select distinct(pk1) from T order by pk1 desc; > +--+ > | PK1 | > +--+ > | 1| > | 2| > | 1| > +--+ > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)