Mailing-List: contact hbase-user-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: hbase-user@hadoop.apache.org
Received-SPF: pass (athena.apache.org: local policy)
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="GB2312"
Content-Transfer-Encoding: quoted-printable
Subject: RE: startRow and endRow doesn't work when use HBase mapreduce
Date: Wed, 23 Dec 2009 18:57:39 +0800
Message-ID: 
 <5A9CF7B9EF3FF34CB589EAC506A1664F03884B@CDCEXMAIL02.tw.trendnet.org>
Thread-Topic: startRow and endRow doesn't work when use HBase mapreduce
Thread-Index: AcqDpe+zTKrpvOoxRRqMy4TOeiWtowAAaqTgAASg+MA=
References: 
 <5A9CF7B9EF3FF34CB589EAC506A1664F038819@CDCEXMAIL02.tw.trendnet.org>
 <61770b880912222359u6b33221ax76b6ed36935881e9@mail.gmail.com> 
From: <Sandy_Yin@trendmicro.com.cn>
To: <hbase-user@hadoop.apache.org>

Hi,

=09I=20tried=20with=20HBase=200.20.2=20,=20the=20startRow=20works=20,=20b=
ut=20the=20endRow=20doesn't=20work=20.=20It=20always=20scan=20to=20the=20=
table=20end.=20

=09I=20reviewed=20the=20source=20code=20of=20HBase=20and=20found=20there=
=20is=20a=20bug=20.

\src\java\org\apache\hadoop\hbase\mapreduce\=20TableInputFormatBase.java=
=20Line=20301

=20298=20=20=20=20=20=20=20byte[]=20splitStart=20=3D=20startRow.length=20=
=3D=3D=200=20||=20
=20299=20=20=20=20=20=20=20=20Bytes.compareTo(keys.getFirst()[i],=20start=
Row)=20>=3D=200=20?=20
=20300=20=20=20=20=20=20=20=20=20=20=20keys.getFirst()[i]=20:=20startRow;
=20301=20=20=20=20=20=20byte[]=20splitStop=20=3D=20stopRow.length=20=3D=
=3D=200=20||=20
=20302=20=20=20=20=20=20=20=20=20Bytes.compareTo(keys.getSecond()[i],=20s=
topRow)=20<=3D=200=20?=20
=20303=20=20=20=20=20=20=20=20=20=20keys.getSecond()[i]=20:=20stopRow;
=20304=20=20=20=20=20=20=20InputSplit=20split=20=3D=20new=20TableSplit(ta=
ble.getTableName(),
=20305=20=20=20=20=20=20=20=20splitStart,=20splitStop,=20regionLocation);

When=20the=20region=20endkey=20is=20empty,=20the=20splitStop=20always=20i=
s=20empty=20,=20So=20the=20endRow=20doesn't=20work.

=09Will=20HBase=200.21=20fix=20this=20bug?

Regards,
Sandy

-----Original=20Message-----
From:=20Sandy=20Yin=20(RD-CN)=20
Sent:=202009=C4=EA12=D4=C223=C8=D5=2016:15
To:=20'hbase-user@hadoop.apache.org'
Subject:=20RE:=20startRow=20and=20endRow=20doesn't=20work=20when=20use=20=
HBase=20mapreduce

Hi=20Lars,

=09Many=20thanks=20for=20the=20information=20:-).
=09I=20tested=20with=20HBase=200.20.1=20,=20it=20doesn't=20work=20.=20I=
=20will=20upgrade=20to=200.20.2=20and=20have=20another=20test.

Regards,
Sandy
=20

-----Original=20Message-----
From:=20Lars=20George=20[mailto:lars.george@gmail.com]=20
Sent:=202009=C4=EA12=D4=C223=C8=D5=2015:59
To:=20hbase-user@hadoop.apache.org
Subject:=20Re:=20startRow=20and=20endRow=20doesn't=20work=20when=20use=20=
HBase=20mapreduce

Hi=20Sandy,

Have=20a=20look=20here:=20http://issues.apache.org/jira/browse/HBASE-1829

I=20added=20tests=20to=20check=20if=20that=20all=20works=20as=20advertise=
d=20and=20it=20indeed
does.=20But=20only=20with=20the=20next=20forthcoming=20versions=20I=20am=
=20afraid.=20With
the=20released=20versions=20I=20would=20have=20assumed=20at=20least=20the=
=20scan=20works
fine=20but=20still=20scans=20the=20whole=20table=20while=20simply=20skipp=
ing=20the=20rows
outside=20the=20given=20range.=20Do=20you=20see=20it=20not=20working=20at=
=20all?

Lars

On=20Wed,=20Dec=2023,=202009=20at=208:46=20AM,=20=20<Sandy_Yin@trendmicro=
.com.cn>=20wrote:
>=20Hi,
>
>
>
>=20The=20startRow=20and=20endRow=20of=20Scan=20doesn't=20work=20when=20u=
se=20HBase=20mapreduce.=20The=20job=20always=20scans=20the=20entire=20tab=
le.
>
>=20Is=20there=20any=20reason=20for=20this=20or=20I=20misuse?
>
>
>
>=20Example=20code:
>
>=20Scan=20scan=20=3D=20new=20Scan();
>
>=20scan.addFamily(...);
>
>=20scan.setStartRow(startkey);
>
>=20scan.setStopRow(endkey);
>
>=20TableMapReduceUtil.initTableMapperJob(tableName,scan,=20mapperClass,=
=20ImmutableBytesWritable.class,Put.class,=20job);
>
>
>
>=20Thanks.
>
>
>=20TREND=20MICRO=20EMAIL=20NOTICE
>=20The=20information=20contained=20in=20this=20email=20and=20any=20attac=
hments=20is=20confidential=20and=20may=20be=20subject=20to=20copyright=20=
or=20other=20intellectual=20property=20protection.=20If=20you=20are=20not=
=20the=20intended=20recipient,=20you=20are=20not=20authorized=20to=20use=
=20or=20disclose=20this=20information,=20and=20we=20request=20that=20you=
=20notify=20us=20by=20reply=20mail=20or=20telephone=20and=20delete=20the=
=20original=20message=20from=20your=20mail=20system.
>

TREND=20MICRO=20EMAIL=20NOTICE
The=20information=20contained=20in=20this=20email=20and=20any=20attachmen=
ts=20is=20confidential=20and=20may=20be=20subject=20to=20copyright=20or=
=20other=20intellectual=20property=20protection.=20If=20you=20are=20not=
=20the=20intended=20recipient,=20you=20are=20not=20authorized=20to=20use=
=20or=20disclose=20this=20information,=20and=20we=20request=20that=20you=
=20notify=20us=20by=20reply=20mail=20or=20telephone=20and=20delete=20the=
=20original=20message=20from=20your=20mail=20system.