Mailing-List: contact dev-help@asterixdb.incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@asterixdb.incubator.apache.org
Sender: Ildar Absalyamov <ty13r.derden@gmail.com>
From: Ildar Absalyamov <ildar.absalyamov@gmail.com>
Content-Type: multipart/alternative;
 boundary="Apple-Mail=_9EDABA8A-68E7-499D-8024-94841655F447"
Message-Id: <56A49DCD-A24A-47DC-AFA7-604A6BFE1B1F@gmail.com>
Mime-Version: 1.0 (Mac OS X Mail 9.1 \(3096.5\))
Subject: Re: Histograms and statistics on Asterix
Date: Fri, 18 Dec 2015 10:18:09 -0800
References: 
 <CALgBV_eZYa+K+H7SKgj3jjMex5w-5vnA0gGGqNXQd-SdoevVZg@mail.gmail.com>
 <5672FBF7.6020504@gmail.com>
 <CALgBV_c4wiWPhpVoRjwFntv-XFsA==6QGgKCwiJe8_Q6twrg9Q@mail.gmail.com>
To: dev@asterixdb.incubator.apache.org
In-Reply-To: 
 <CALgBV_c4wiWPhpVoRjwFntv-XFsA==6QGgKCwiJe8_Q6twrg9Q@mail.gmail.com>

--Apple-Mail=_9EDABA8A-68E7-499D-8024-94841655F447
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii

Hi Wail,

I do have some doc which can give you some idea of how stats collection =
will work =
https://docs.google.com/document/d/12aP8Pzp68b_HxJe-svFmG9dHFX6zliGw-ErNSd=
_Av3I/edit?usp=3Dsharing =
<https://docs.google.com/document/d/12aP8Pzp68b_HxJe-svFmG9dHFX6zliGw-ErNS=
d_Av3I/edit?usp=3Dsharing>. But it does not describe the statistical =
synopsis format. We were planning to use wavelets instead to histograms =
for these LSM-based statistics. Let me know if you want to know more on =
how wavelets work and why they should be used.

> On Dec 18, 2015, at 07:23, Wail Alkowaileet <wael.y.k@gmail.com> =
wrote:
>=20
> Good to know.
> Is there any type of design document ?
>=20
> P.S. Sattam told me that there was some sort of a comparative study =
shows
> the difference in performance of open vs. closed types. Where can I =
find it?
>=20
> Thanks!
>=20
> On Thu, Dec 17, 2015 at 9:16 PM, Mike Carey <dtabass@gmail.com> wrote:
>=20
>> Ildar is working on that (camping on the LSM lifecycle) and Wenhai is
>> doing some work on histograms that's intended for in-flight use =
during
>> query processing (to do dynamic partitioning, e.g., during parallel =
sorts
>> and joins).
>>=20
>>=20
>> On 12/17/15 1:53 AM, Wail Alkowaileet wrote:
>>=20
>>> Hi Devs,
>>>=20
>>> I want to ask if there's any on going work on building histograms =
and
>>> stats
>>> for Asterix ?
>>>=20
>>=20
>>=20
>=20
>=20
> --=20
>=20
> *Regards,*
> Wail Alkowaileet

Best regards,
Ildar


--Apple-Mail=_9EDABA8A-68E7-499D-8024-94841655F447--