Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (nike.apache.org: domain of filippo@ntoklo.com designates
 209.85.215.172 as permitted sender)
Date: Tue, 22 May 2012 13:03:58 +0100
From: Filippo Diotalevi <filippo@ntoklo.com>
To: user@cassandra.apache.org
Message-ID: <520F8B47B1EE422781E36EFA6443C176@ntoklo.com>
In-Reply-To: 
 <OFADD130B9.D530A06C-ONC1257A06.002AE2B9-C1257A06.002AFA12@urssaf.fr>
References: 
 <OFADD130B9.D530A06C-ONC1257A06.002AE2B9-C1257A06.002AFA12@urssaf.fr>
Subject: Re: RE Ordering counters in Cassandra
MIME-Version: 1.0
Content-Type: multipart/alternative; boundary="4fbb80ae_5451cf49_424"

--4fbb80ae_5451cf49_424
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline

Thanks for all the answers, they definitely helped. =20

Just out of curiosity, is there any underlying architectural reason why i=
t's not possible to order a row based on its counters values=3F or is it =
something that might be in the roadmap in the future=3F =20

-- =20
=46ilippo Diotalevi


On Tuesday, 22 May 2012 at 08:48, Romain HARDOUIN wrote:

> =20
> I mean iterate over each column -- more precisly: *bunches of columns* =
using slices -- and write new columns in the inversed index. =20
> Tamar's data model is made for real time analysis. It's maybe overdesig=
ned for a daily ranking. =20
> I agree with Samal, you should split your data across the space of toke=
ns. Only C=46 Ranking feeding would be affected, not the =22top N=22 quer=
ies. =20
> =20
> =46ilippo Diotalevi <filippo=40ntoklo.com (mailto:filippo=40ntoklo.com)=
> a =C3=A9crit sur 21/05/2012 19:05:28 :
> =20
> > Hi Romain, =20
> > thanks for your suggestion. =20
> > =20
> > When you say =22 build every day a ranking in a dedicated C=46 by =20
> > iterating over events:=22 do you mean =20
> > - load all the columns for the specified row key =20
> > - iterate over each column, and write a new column in the inversed in=
dex =20
> > =3F =20
> > =20
> > That's my current approach, but since I have many of these wide rows
> > (1 per day), the process is extremely slow as it involves moving an =20
> > entire row from Cassandra to client, inverting every column, and =20
> > sending the data back to create the inversed index. =20


--4fbb80ae_5451cf49_424
Content-Type: text/html; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline


                <div>
                    Thanks for all the answers, they definitely helped.
                </div><div><br></div><div>Just out of curiosity, is there=
 any underlying architectural reason why it's not possible to order a row=
 based on its counters values=3F or is it something that might be in the =
roadmap in the future=3F</div>
                <div><div><br></div><div>--&nbsp;</div><div>=46ilippo Dio=
talevi</div><div><br></div></div>
                =20
                <p style=3D=22color: =23A0A0A8;=22>On Tuesday, 22 May 201=
2 at 08:48, Romain HARDOUIN wrote:</p>
                <blockquote type=3D=22cite=22 style=3D=22border-left-styl=
e:solid;border-width:1px;margin-left:0px;padding-left:10px;=22>
                    <span><div><div>
<br><font size=3D=222=22 face=3D=22sans-serif=22>I mean iterate over each=
 column -- more
precisly: *bunches of columns* using slices -- and write new columns in
the inversed index.</font>
<br><font size=3D=222=22 face=3D=22sans-serif=22>Tamar's data model is ma=
de for real
time analysis. It's maybe overdesigned for a daily ranking.</font>
<br><font size=3D=222=22 face=3D=22sans-serif=22>I agree with Samal, you =
should split
your data across the space of tokens. Only C=46 Ranking feeding would be
affected, not the =22top N=22 queries.</font>
<br>
<br><font size=3D=222=22><tt>=46ilippo Diotalevi &lt;<a href=3D=22mailto:=
filippo=40ntoklo.com=22>filippo=40ntoklo.com</a>&gt; a =C3=A9crit
sur 21/05/2012 19:05:28 :<br>
<br>
&gt; Hi Romain, </tt></font>
<br><font size=3D=222=22><tt>&gt; thanks for your suggestion.</tt></font>=

<br><font size=3D=222=22><tt>&gt; <br>
&gt; When you say =22 build every day a ranking in a dedicated C=46 by
<br>
&gt; iterating over events:=22 do you mean</tt></font>
<br><font size=3D=222=22><tt>&gt; - load all the columns for the specifie=
d row
key</tt></font>
<br><font size=3D=222=22><tt>&gt; - iterate over each column, and write a=
 new column
in the inversed index</tt></font>
<br><font size=3D=222=22><tt>&gt; =3F</tt></font>
<br><font size=3D=222=22><tt>&gt; <br>
&gt; That's my current approach, but since I have many of these wide rows=
<br>
&gt; (1 per day), the process is extremely slow as it involves moving an
<br>
&gt; entire row from Cassandra to client, inverting every column, and <br=
>
&gt; sending the data back to create the inversed index.</tt></font>
<br></div></div></span>
                =20
                =20
                =20
                =20
                </blockquote>
                =20
                <div>
                    <br>
                </div>
            
--4fbb80ae_5451cf49_424--