[ https://issues.apache.org/jira/browse/HIVE-13312?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15288866#comment-15288866 ]
Pau Tallada CrespĂ commented on HIVE-13312: ------------------------------------------- Also, even with the workaround, an aggregated query reads and processes all rows. So, a count(*) on a tablesample returns the actual number of rows in the entire table :( > TABLESAMPLE with PERCENT throws FAILED: SemanticException 1:68 Percentage > sampling is not supported in org.apache.hadoop.hive.ql.io.HiveInputFormat. > Error encountered near token '20' > -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- > > Key: HIVE-13312 > URL: https://issues.apache.org/jira/browse/HIVE-13312 > Project: Hive > Issue Type: Bug > Components: CLI > Affects Versions: 1.2.1 > Reporter: Artem Ervits > Priority: Minor > > FAILED: SemanticException 1:68 Percentage sampling is not supported in > org.apache.hadoop.hive.ql.io.HiveInputFormat. Error encountered near token > '20' > when I execute > SELECT * FROM tablename TABLESAMPLE(20 percent); > tried with ORC and TEXT tables. Confirmed with Gopal, a temporary workaround > is > set hive.tez.input.format=${hive.input.format}; -- This message was sent by Atlassian JIRA (v6.3.4#6332)