[ https://issues.apache.org/jira/browse/MAPREDUCE-2614?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13053931#comment-13053931 ]
Josh Wills commented on MAPREDUCE-2614: --------------------------------------- I agree that the ability to use ON DUPLICATE KEY UPDATE syntax for outputs that write to MySQL would be useful-- as would the ability to specify INSERT IGNORE and turn speculative reduce execution back on if you knew that your output table had unique keys. But I also imagine that Oracle users would like to be able to use INSERT WHEN syntax, or use Oracle's TO_DATE function on some inserts. So I think that the right way to solve this problem is to decouple the query building logic from DBRecordWriter- create a generic default impl that does the same thing as the current code, and then create MySQL/Oracle/etc. specific implementations of the query building interface. > Allow append arbitrary text at the end of generated query in DBOutputFormat > class > --------------------------------------------------------------------------------- > > Key: MAPREDUCE-2614 > URL: https://issues.apache.org/jira/browse/MAPREDUCE-2614 > Project: Hadoop Map/Reduce > Issue Type: Improvement > Reporter: Jarek Jarcec Cecho > Priority: Minor > Attachments: MAPREDUCE-2614.patch > > > It would be wonderful if DBOutputFormat class allow addition of arbitrary > text at the end of generated query. This feature can be useful for example in > case of MySQL database to specify "ON DUPLICATE KEY UPDATE ..." part of the > query. -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira