Mailing-List: contact user-help@hive.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hive.apache.org
Received-SPF: pass (nike.apache.org: domain of
 Sanjay.Subramanian@wizecommerce.com designates 207.46.163.25 as permitted
 sender)
Received-SPF: pass (mail205-co9: domain of wizecommerce.com designates
 157.56.232.197 as permitted sender) client-ip=157.56.232.197;
 envelope-from=Sanjay.Subramanian@wizecommerce.com;
 helo=BLUPRD0411HT004.namprd04.prod.outlook.com ;.outlook.com ;
From: Sanjay Subramanian <Sanjay.Subramanian@wizecommerce.com>
To: "user@hive.apache.org" <user@hive.apache.org>
Subject: Re: Update statment on Hive
Thread-Topic: Update statment on Hive
Thread-Index: AQHOXjUNRp0ehlVPsk66SJymlhh8mZkfQkCA
Date: Fri, 31 May 2013 20:04:08 +0000
Message-ID: <CDCE4EE7.7E8B%sanjay.subramanian@wizecommerce.com>
In-Reply-To: <1370028452.24112.5.camel@oc5123520263.ibm.com>
Accept-Language: en-US
Content-Language: en-US
Content-Type: text/plain; charset="Windows-1252"
Content-ID: <BD05027E506D494E9E98164275EBB012@namprd04.prod.outlook.com>
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0

Hi
Hive reads and writes to HDFS=8Aand by definition HDFS is write once and
immutable after that.
So like an RDBMS there is no concept of an update rows.
However if u want to delete some records based on a criteria, yesterday
there was a smart post about it, basically selecting the inverse and doing
an INSERT OVERWRITE on the table

INSERT OVERWRITE TABLE will write to the Hive managed HDFS location of
that table and "replace" all that is there with your latest

INSERT OVERWRITE DIRECTORY will write to your HDFS location of choice and
"replace" all that is there with your latest. You can use this directory
as LOCATION to which your PARTITION may point later


Note that INSERT OVERWRITE TABLE will follow the field separator of the
destination table that u specified while creating the table

However INSERT OVERWRITE DIRECTORY will have Hive's default "CtrlA" as the
field separator ( I use 0.10.x)=8Aperhaps this is changed in 0.11 u need to
verify

In summary , u have to look at updates and deletes very differently from
RDBMS

Good luck

sanjay


On 5/31/13 12:27 PM, "Renata Ghisloti Duarte de Souza"
<rgduarte@linux.vnet.ibm.com> wrote:

>Hello,
>
>I was wondering about the "update" statement on Hive. Is it something
>Hive needs? Or can "insert overwrite" be always used instead?
>
>Thank you in advance for the clarification,
>
>Renata.
>
>
>


CONFIDENTIALITY NOTICE
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
This email message and any attachments are for the exclusive use of the int=
ended recipient(s) and may contain confidential and privileged information.=
 Any unauthorized review, use, disclosure or distribution is prohibited. If=
 you are not the intended recipient, please contact the sender by reply ema=
il and destroy all copies of the original message along with any attachment=
s, from your computer system. If you are the intended recipient, please be =
advised that the content of this message is subject to access, review and d=
isclosure by the sender's Email System Administrator.