Keep in mind that if your use case is to guarantee an empty table to load into then every 'insert overwrite table' does just that for you (it clobbers vs insert into). This and the workaround (which must be table specific) is complicated when you have partitions. If your usecase is reclaiming space, just data invalidation, or have partitions simply deleting the base data via hadoop -rmr /user/hive/warehouse/<table>/* (assuming default data mgmt) is decoupled but works and should not be harmful -- Hive is supposed to be just an overlay but there was a change to partition mgmt that I would like to hear whether people think raw delete is now undesirable.
On Apr 26, 2009 12:34 AM, "Ashish Thusoo" <[email protected]> wrote: Suhail, Hive does not support truncate yet. I think Prasad already opened a JIRA for this https://issues.apache.org/jira/browse/HIVE-446 Matt, I must say that this is a very creative workaround :) Ashish ________________________________________ From: Matt Pestritto [[email protected]] Sent: Saturday, April 25, 2009 10:08 AM To: [email protected] Subject: Re: Truncate Table I've used the following: Insert overwrite tbl select * from tbl where col1 = 'fake value' On 4/25/...
