[ https://issues.apache.org/jira/browse/HIVE-2987?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13267073#comment-13267073 ]
Navis commented on HIVE-2987: ----------------------------- It seemed not possible to discern empty row from not-existing row in hbase. > SELECTing nulls returns nothing > ------------------------------- > > Key: HIVE-2987 > URL: https://issues.apache.org/jira/browse/HIVE-2987 > Project: Hive > Issue Type: Bug > Affects Versions: 0.9.0 > Environment: Tested using 0.9.0rc1, hbase 0.92.1, hadoop 0.20.2-cdh3u2 > Reporter: Oliver Meyn > Priority: Critical > > Given an hbase table defined as 'test' with a single column family 'a', > rowkey of type string, and two "rows" as follows: > key:1,a:lat=60.0,a:long=50.0,a:precision=10 > key:2,a:lat=54 > And an hive table created overtop of it as follows: > CREATE EXTERNAL TABLE hbase_test ( > id STRING, > latitude STRING, > longitude STRING, > precision STRING > ) > STORED BY 'org.apache.hadoop.hive.hbase.HBaseStorageHandler' > WITH SERDEPROPERTIES ("hbase.columns.mapping" = > ":key#s,a:lat#s,a:long#s,a:precision#s") > TBLPROPERTIES( > "hbase.table.name" = "test", > "hbase.table.default.storage.type" = "binary" > ); > The query SELECT id, precision FROM hbase_test WHERE id = '2' returns no > result. Expected behaviour is to return: > '2',NULL > If the query is changed to include a non-null result, eg SELECT id, latitude, > precision FROM hbase_test WHERE id = '2' the result is as expected: > '2','54',NULL -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira