Mailing-List: contact user-help@cassandra.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@cassandra.apache.org
Received-SPF: pass (athena.apache.org: domain of augustyn.michal@gmail.com
 designates 209.85.216.44 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:date:message-id:subject:from:to
         :content-type;
        b=K8Enq0+t6N0YoaBgCnT4iSfj3bk3cXtbxmrIi1AVWde4qla1Vs/PwgauwD995dBX4E
         gaPcBX+xPMdXpQ4myc4ZnOEUl0MgE4GrTBe5AyeB29dvTeIhTug0rPI/ZpmHgxde7uIK
         /H7LraFyXHN+1k/HfC0u7i4vhmo0AMpuR+m90=
MIME-Version: 1.0
In-Reply-To: <1283710474.639320410@192.168.2.229>
References: <AANLkTi=CTtigBCSOoNvbUK2GFR=GO4pkrVFCi8iup5-r@mail.gmail.com>
	<1283710474.639320410@192.168.2.229>
Date: Sun, 5 Sep 2010 21:30:49 +0200
Message-ID: <AANLkTinf1Ym2kCHEKaEk7NdB8pHX-QLGcXoavYA=6KUS@mail.gmail.com>
Subject: Re: skip + limit support in GetSlice
From: =?ISO-8859-1?Q?Michal_August=FDn?= <augustyn.michal@gmail.com>
To: user@cassandra.apache.org
Content-Type: multipart/alternative; boundary=0015175cda328aea58048f88332a

--0015175cda328aea58048f88332a
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

I know that "Prev/Next" is good solution for web applications. But when I
want to access data from another application or when I want to access pages
randomly...

I don't know the internal structure of memtables etc., so I don't know if
columns in row are indexable. If now, then I just want to transfer my
workaround to server (to avoid huge network traffic)...

2010/9/5 Stu Hood <stu.hood@rackspace.com>

> Cassandra supports the recommended approach from:
> http://www.percona.com/ppc2009/PPC2009_mysql_pagination.pdf
>
> For large numbers of items, skip + limit is extremely inefficent.
>
> -----Original Message-----
> From: "Michal August=FDn" <augustyn.michal@gmail.com>
> Sent: Sunday, September 5, 2010 5:39am
> To: user@cassandra.apache.org
> Subject: skip + limit support in GetSlice
>
> Hello,
>
> probably this is feature request. Simply, I would like to have support fo=
r
> standard pagination (skip + limit) in GetSlice Thrift method. Is this
> feature on the road map?
>
> Now, I have to perform GetSlice call, that starts on "" and "limit" is se=
t
> to "skip" value. Then I read the last column name returned and subsequent=
ly
> perform the final GetSlice call - I use the last column name as "start" a=
nd
> set "limit" to "limit" value.
>
> This workaround is not very efficient when I need to skip a lot of column=
s
> (so "skip" is high) - then a lot of data must be transferred via network.
> So
> I think that support for Skip in GetSlice would be very useful (to avoid
> high network traffic).
>
> The implementation could be very straightforward (same as the workaround)
> or
> maybe it could be more efficient - I think that whole row (so all columns=
)
> must fit into memory so if we have all columns in memory...
>
> Thank you!
>
> Augi
>
>
>

--0015175cda328aea58048f88332a
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

I know that &quot;Prev/Next&quot; is good solution for web applications. Bu=
t when I want to access data from another application or when I want to acc=
ess pages randomly...<div><br></div><div>I don&#39;t know the internal stru=
cture of memtables etc., so I don&#39;t know if columns in row are=A0indexa=
ble. If now, then I just want to transfer my workaround to server (to avoid=
 huge network traffic)...<br>
<br><div class=3D"gmail_quote">2010/9/5 Stu Hood <span dir=3D"ltr">&lt;<a h=
ref=3D"mailto:stu.hood@rackspace.com">stu.hood@rackspace.com</a>&gt;</span>=
<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-lef=
t:1px #ccc solid;padding-left:1ex;">
Cassandra supports the recommended approach from: <a href=3D"http://www.per=
cona.com/ppc2009/PPC2009_mysql_pagination.pdf" target=3D"_blank">http://www=
.percona.com/ppc2009/PPC2009_mysql_pagination.pdf</a><br>
<br>
For large numbers of items, skip + limit is extremely inefficent.<br>
<div><div></div><div class=3D"h5"><br>
-----Original Message-----<br>
From: &quot;Michal August=FDn&quot; &lt;<a href=3D"mailto:augustyn.michal@g=
mail.com">augustyn.michal@gmail.com</a>&gt;<br>
Sent: Sunday, September 5, 2010 5:39am<br>
To: <a href=3D"mailto:user@cassandra.apache.org">user@cassandra.apache.org<=
/a><br>
Subject: skip + limit support in GetSlice<br>
<br>
Hello,<br>
<br>
probably this is feature request. Simply, I would like to have support for<=
br>
standard pagination (skip + limit) in GetSlice Thrift method. Is this<br>
feature on the road map?<br>
<br>
Now, I have to perform GetSlice call, that starts on &quot;&quot; and &quot=
;limit&quot; is set<br>
to &quot;skip&quot; value. Then I read the last column name returned and su=
bsequently<br>
perform the final GetSlice call - I use the last column name as &quot;start=
&quot; and<br>
set &quot;limit&quot; to &quot;limit&quot; value.<br>
<br>
This workaround is not very efficient when I need to skip a lot of columns<=
br>
(so &quot;skip&quot; is high) - then a lot of data must be transferred via =
network. So<br>
I think that support for Skip in GetSlice would be very useful (to avoid<br=
>
high network traffic).<br>
<br>
The implementation could be very straightforward (same as the workaround) o=
r<br>
maybe it could be more efficient - I think that whole row (so all columns)<=
br>
must fit into memory so if we have all columns in memory...<br>
<br>
Thank you!<br>
<br>
Augi<br>
<br>
<br>
</div></div></blockquote></div><br></div>

--0015175cda328aea58048f88332a--