Mailing-List: contact user-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hbase.apache.org
MIME-Version: 1.0
In-Reply-To: <CAHu1D0hFDg4NqcipCEo5Esj4yFUFodqkNJ4ueWuLdw-x8ei+ig@mail.gmail.com>
References: <CAHu1D0iqDj4WpSgWtL_=3HK-6eObXZvkSzFp7FdMLkbeO1xERA@mail.gmail.com>
 <CALte62yuLnAzHFZH9-hDLV-=JBuUT_=ednqYP4uHWB9vCMpBBA@mail.gmail.com> <CAHu1D0hFDg4NqcipCEo5Esj4yFUFodqkNJ4ueWuLdw-x8ei+ig@mail.gmail.com>
From: Ted Yu <yuzhihong@gmail.com>
Date: Wed, 17 May 2017 06:28:31 -0700
Message-ID: <CALte62yi7Eq=6Y68ciFiVXUh36rxxwto1QEsydBiGN5sNXizgg@mail.gmail.com>
Subject: Re: hbase coprocessor in mapreduce
To: "user@hbase.apache.org" <user@hbase.apache.org>
Content-Type: multipart/alternative; boundary="001a113edee4b64dff054fb84395"
archived-at: Wed, 17 May 2017 13:28:43 -0000

--001a113edee4b64dff054fb84395
Content-Type: text/plain; charset="UTF-8"

You can take a look at the following from Phoenix:
phoenix-core/src/main/java/org/apache/phoenix/coprocessor/GroupedAggregateRegionObserver.java

On Wed, May 17, 2017 at 1:59 AM, Rajeshkumar J <rajeshkumarit8292@gmail.com>
wrote:

> Hi,
>
>   I want to groupby hbase data by some columns and again I want to group
> the resultant data by some columns and check the size of each group if
> greater than 1 it will be written to hdfs. Initially we are doing this in
> mapreduce but we got some issue in reducer. So we thought we can move this
> group by operation to hbase using coprocessor.
>   Hence using this coprocessor we will scan and group data by some columns
> . Then again I want to group the resultant data how to achieve this in
> coprocessor?? whether this use case fits in coprocessor??
>
> Thanks
>
> On Wed, May 17, 2017 at 1:28 PM, Ted Yu <yuzhihong@gmail.com> wrote:
>
> > Can you clarify your use case ?
> >
> > If your mapper calls Table#coprocessorService(), the recipient may not be
> > on the same region server as the mapper.
> >
> > On Tue, May 16, 2017 at 10:58 PM, Rajeshkumar J <
> > rajeshkumarit8292@gmail.com
> > > wrote:
> >
> > > Hi,
> > >
> > >    whether I can call hase coprocessor in the map function of mapreduce
> > > job?
> > >
> > > Thanks
> > >
> >
>

--001a113edee4b64dff054fb84395--