Mailing-List: contact dev-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@hbase.apache.org
Received-SPF: pass (athena.apache.org: local policy)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=yahoo.com;
  h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-RocketYMMF:X-Mailer:References:Message-ID:Date:From:Reply-To:Subject:To:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding;
  b=jL8af4oVM4ifCdY2pLo279BwPE5FGyadPpnY44nZYP7xJ4+a+ml85DjA8CU8GpY+X0LQIXaRl+RPQBGxSY5VL1DiVBTo73ah/yJjx5uIN+276wAhIXmJfvyZwTky3pbKZPNqRloU/SXEytJ1LOMOW5ccGiO4yHq3R0daJzI9JO4=;
References: 
 <CAAg3a2oOTyoSCRGE5poUu=32NyB2iq1Wx3XYx0MRPfctFroTPA@mail.gmail.com>
 <CADcMMgGi05E1KY+wvWAdZJPr6ddH6zR0KU110gjf0pyvaUP0ew@mail.gmail.com>
Message-ID: <1381619853.56048.YahooMailNeo@web140605.mail.bf1.yahoo.com>
Date: Sat, 12 Oct 2013 16:17:33 -0700 (PDT)
From: lars hofhansl <larsh@apache.org>
Reply-To: lars hofhansl <larsh@apache.org>
Subject: Re: Excessive filtering in RegionScannerImpl
To: "dev@hbase.apache.org" <dev@hbase.apache.org>
In-Reply-To: 
 <CADcMMgGi05E1KY+wvWAdZJPr6ddH6zR0KU110gjf0pyvaUP0ew@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable

Filters can work at the row level and the KV level.=0AIn RegionScannerImpl.=
nextInternal we do the row level filtering and in StoreScanner.next() via S=
canQueryMatcher we do the KV based filtering.=0ASo I would not call this ex=
cessive, but rather that we are doing the filtering at the appropriate leve=
l. I'm sure that can be improved perf wise, though.=0A=0A-- Lars=0A=0A=0A=
=0A----- Original Message -----=0AFrom: Stack <stack@duboce.net>=0ATo: HBas=
e Dev List <dev@hbase.apache.org>=0ACc: =0ASent: Saturday, October 12, 2013=
 1:55 PM=0ASubject: Re: Excessive filtering in RegionScannerImpl=0A=0AOn Sa=
t, Oct 12, 2013 at 11:25 AM, Vladimir Rodionov=0A=0A<vladrodionov@gmail.com=
>wrote:=0A=0A> I may be be do not understand fully the overall design of a =
scanner system=0A> in HBase but it looks like we do double filtering on KVs=
 during scan=0A> operation. First time in StoreScanner.next (see ScanQueryM=
atcher.match) and=0A> second time in RegionScannerImpl.nextInternal.=0A>=0A=
=0A=0AQuiet possible.=A0 LarsH?=0AThanks for looking into this Vladimir.=0A=
St.Ack=0A