Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: pass (nike.apache.org: local policy includes SPF record at
 spf.trusted-forwarder.org)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=sbcglobal.net;
  h=X-YMail-OSG:Received:X-Rocket-MIMEInfo:X-Mailer:References:Message-ID:Date:From:Reply-To:Subject:To:In-Reply-To:MIME-Version:Content-Type;
  b=tJmJomoYL4lvM9oRaBZ9qdnJtgQeovch0JSP7lkmX7iVRtgsryLAWaDveo6Xmuzrj35IFzbUgRwJHCBsmpJnUE7IMty5JR8Wtm6rqiVra2JdPvm758Fd8Kkj43vG40jgvRSI4aL6ToOMylAJQl7T0mWtI5YXkuimVlmoqmPqwcU=;
References: <0F9923A3-3054-4695-A66B-F8A3E0FD475D@apache.org>
 <2CEABFBA-FBE9-42CC-817F-24A56B6DF36E@yahoo.com>
 <d594246002de4de6b867f9ac2040a659@BY2PR05MB253.namprd05.prod.outlook.com>
Message-ID: <1396529627.10521.YahooMailNeo@web181702.mail.ne1.yahoo.com>
Date: Thu, 3 Apr 2014 05:53:47 -0700 (PDT)
From: Scott Weber <scotty2541@sbcglobal.net>
Reply-To: Scott Weber <scotty2541@sbcglobal.net>
Subject: Re: question about "complex" range queries
To: COUCHDB <user@couchdb.apache.org>
In-Reply-To: 
 <d594246002de4de6b867f9ac2040a659@BY2PR05MB253.namprd05.prod.outlook.com>
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary="-1672125993-556949270-1396529627=:10521"

---1672125993-556949270-1396529627=:10521
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable

I have been following this, and was wondering about it.=0A=0AThe documentat=
ion doesn't say anything about the second arg (i.e. value) of the emit call=
 being used for sorting.=C2=A0 However it works that way every time, and I =
can find no situation where it doesn't happen.=0A=0AAre you saying that it =
is NOT something that should be relied on?=0A=0AI also notice that the docu=
ments are sent into the view already sorted by their _id.=C2=A0 Is that a b=
ehavior that CAN be replied on?=0A=0ASee my samples below:=0A=0A=0ATwo of m=
y docs:=0A{=C2=A0=C2=A0 "_id": "a",=C2=A0=C2=A0 "_rev": "1-a50d41b0fdfaaca4=
e679c7f2c6a1722d",=0A=C2=A0=C2=A0 "items": [=C2=A0=C2=A0 =0A=0A=C2=A0 { "b"=
: "5"}, {"a": "6"}, {"c": "4"}, { "x": "1" }, { "w": "3" }, { "z": "2" }=0A=
]=C2=A0 }=0A=0A=0A{=C2=A0 "_id": "b", "_rev": "1-82a40a986a557f05352dbcb1fb=
07b26e",=0A=C2=A0=C2=A0 "items": [=0A=C2=A0=C2=A0 {"m": "5" },{ "l": "6" },=
 { "n": "4"}=0A] }=0A=0A=0AThe temp view:=0Afunction(doc) {=C2=A0=C2=A0=C2=
=A0 var x;=0Afor (x =3D 0; x < doc.items.length; x++) {=C2=A0 emit("1", doc=
.items[x]); =C2=A0 }=0A}=0A=0A=0AThe output:=C2=A0 =0A=0A{"total_rows":9,"o=
ffset":0,"rows":[=0A{"id":"a","key":"1","value":{"a":"6"}},=0A{"id":"a","ke=
y":"1","value":{"b":"5"}},=0A{"id":"a","key":"1","value":{"c":"4"}},=0A{"id=
":"a","key":"1","value":{"w":"3"}},=0A{"id":"a","key":"1","value":{"x":"1"}=
},=0A{"id":"a","key":"1","value":{"z":"2"}},=0A{"id":"b","key":"1","value":=
{"l":"6"}},=0A{"id":"b","key":"1","value":{"m":"5"}},=0A{"id":"b","key":"1"=
,"value":{"n":"4"}}=0A]}=0A=0A=0A-Scott=0A=0A=0A=0A________________________=
________=0A=0AFrom: Robert Samuel Newson <rnewson@apache.org>=0A>Date: Apri=
l 3, 2014, 3:30:14 AM CDT=0A>To: user@couchdb.apache.org, "Rian R. Maloney"=
 <rian.maloney@yahoo.com>=0A>Subject: Re: question about "complex" range qu=
eries=0A>Reply-To: user@couchdb.apache.org=0A=0A>Sorry, but the value is no=
t included in the ordering of the views, what you=E2=80=99re seeing there i=
s an artifact of how multiple emits from the same document (where the emitt=
ed key is the same *and* the tie-breaking _id is also the same) are handled=
.=0A>=0A>B.=0A>=0A>On 3 Apr 2014, at 01:52, Rian R. Maloney <rian.maloney@y=
ahoo.com> wrote:=0A>=0A>=0A>=0A>This is NOT what I am seeing ( just trying =
to understand how this example is not sorted on value). Here is my test cas=
e:=0A>=C2=A0=0A>1) create database called test=0A>2) create 1 doc with _id =
=3D test and a field called stuff =3D"this is test data"=0A>3) =C2=A0create=
 the following map:=0A>=C2=A0=0A>function(doc) {=0A>=C2=A0=0A>=C2=A0for( va=
r i =3D 1;i < 5;i++ )=0A>=C2=A0{=0A>=C2=A0=C2=A0=C2=A0=C2=A0data =3D {};=0A=
>=C2=A0=C2=A0=C2=A0=C2=A0data.sortkey =3D i;=0A>=C2=A0=C2=A0=C2=A0=C2=A0dat=
a.stuff =3D doc.stuff;=0A>=C2=A0=C2=A0=C2=A0=C2=A0emit( doc._id, data );=0A=
>=C2=A0}=0A>=C2=A0for( var i =3D 4;i > 0;i-- )=0A>=C2=A0{=0A>=C2=A0=C2=A0=
=C2=A0=C2=A0data =3D {};=0A>=C2=A0=C2=A0=C2=A0=C2=A0data.sortkey =3D i;=0A>=
=C2=A0=C2=A0=C2=A0=C2=A0data.stuff =3D doc.stuff;=0A>=C2=A0=C2=A0=C2=A0=C2=
=A0emit( doc._id, data );=0A>=C2=A0}=0A>=C2=A0=0A>=C2=A0=0A>Result:=0A>{"to=
tal_rows":8,"offset":0,"rows":[=0A>{"id":"test","key":"test","value":{"sort=
key":1,"stuff":"this is test data"}},=0A>{"id":"test","key":"test","value":=
{"sortkey":1,"stuff":"this is test data"}},=0A>{"id":"test","key":"test","v=
alue":{"sortkey":2,"stuff":"this is test data"}},=0A>{"id":"test","key":"te=
st","value":{"sortkey":2,"stuff":"this is test data"}},=0A>{"id":"test","ke=
y":"test","value":{"sortkey":3,"stuff":"this is test data"}},=0A>{"id":"tes=
t","key":"test","value":{"sortkey":3,"stuff":"this is test data"}},=0A>{"id=
":"test","key":"test","value":{"sortkey":4,"stuff":"this is test data"}},=
=0A>{"id":"test","key":"test","value":{"sortkey":4,"stuff":"this is test da=
ta"}}=0A>]}=0A>=C2=A0=0A>But if value was not included in the sort, it woul=
d be:=0A>=C2=A0=0A>{"total_rows":8,"offset":0,"rows":[=0A>{"id":"test","key=
":"test","value":{"sortkey":1,"stuff":"this is test data"}},=0A>{"id":"test=
","key":"test","value":{"sortkey":2,"stuff":"this is test data"}},=0A>{"id"=
:"test","key":"test","value":{"sortkey":3,"stuff":"this is test data"}},=0A=
>{"id":"test","key":"test","value":{"sortkey":4,"stuff":"this is test data"=
}},{"id":"test","key":"test","value":{"sortkey":1,"stuff":"this is test dat=
a"}},{"id":"test","key":"test","value":{"sortkey":2,"stuff":"this is test d=
ata"}},{"id":"test","key":"test","value":{"sortkey":3,"stuff":"this is test=
 data"}},{"id":"test","key":"test","value":{"sortkey":4,"stuff":"this is te=
st data"}},=0A>]}=0A>On Wednesday, April 2, 2014 6:07 PM, Ryan Ramage <ryan=
.ramage@gmail.com> wrote:=0A>=C2=A0=0A>The emitted value does not affect th=
e sort.=0A>=C2=A0=0A>=C2=A0=0A>If you wanted to keep sorting add your sort =
field in an array key, so move=0A>from=0A>=C2=A0=0A>emit(doc.key, doc.value=
)=0A>=C2=A0=0A>=C2=A0=0A>to=0A>=C2=A0=0A>emit([doc.key, doc.value], doc.val=
ue)=0A>=C2=A0=0A>=C2=A0=0A>=C2=A0=0A>On Wed, Apr 2, 2014 at 4:57 PM, <rian.=
maloney@yahoo.com> wrote:=0A>=C2=A0=0A>Thanks Robert. Let me clarify. Is th=
e value included in the sort? I have=0A>10 entries all with identical keys.=
 The behavior I am seeing indicates the=0A>contents of the emitted value is=
 included in the sort criteria and if so=0A>can I count on this? Bottom lin=
e I am trying to sort on 2 fields, one is a=0A>key and the other is the beg=
inning of the value. I also don't want to sort=0A>the results in a list due=
 to performance constraints=0A>=C2=A0=0A>Thanks=0A>=C2=A0=0A>On Apr 2, 2014=
, at 5:45 PM, Robert Samuel Newson <rnewson@apache.org>=0A>wrote:=0A>=C2=A0=
=0A>A view is a single dimension index, ordered by the full key in=0A>accor=
dance with the rules enumerated here:=0A>https://wiki.apache.org/couchdb/Vi=
ew_collation#Collation_Specification=0A>=C2=A0=0A>For example;=0A>=C2=A0=0A=
>9 sorts after false=0A>999 sorts after 9=0A>"foo" sorts after 999=0A>["foo=
"] sorts after "foo"=0A>["foo", "bar"] sorts after ["foo"]=0A>=C2=A0=0A>B.=
=0A>=C2=A0=0A>On 2 Apr 2014, at 23:29, rian.maloney@yahoo.com wrote:=0A>=C2=
=A0=0A>I have a quick question related to this. What exactly is sorted? My=
=0A>test cases indicated the key AND the entire value is sorted as if it wa=
s=0A>one big string.=0A>=C2=A0=0A>Thanks=0A>Rian=0A>=C2=A0=0A>On Apr 2, 201=
4, at 2:31 AM, Nick North <north.n@gmail.com> wrote:=0A>=C2=A0=0A>It doesn'=
t work quite like this. There is a single order across all=0A>possible keys=
, including both simple and complex keys, as described=0A>here<https://wiki=
.apache.org/couchdb/View_collation>.=0A>In the case of keys that are lists,=
 the two lists are compared element=0A>by=0A>element and the sort order is =
the sort order of the first unequal=0A>elements.=0A>=C2=A0=0A>In your examp=
le, if a key has its first element between a1 and a2 (and=0A>a1=0A>and a2 a=
re different), then the second element will not be inspected at=0A>all, so =
it does not matter whether it is between b1 and b2 or not. In=0A>fact=0A>th=
e second element will only be inspected if the first element is=0A>either a=
1=0A>or a2.=0A>=C2=A0=0A>This is usually the behaviour we want. For example=
, dates are often=0A>represented as lists of [year, month, day]. Then you c=
an pull out all=0A>the=0A>documents in a date range by specifying start dat=
e and end date as=0A>startkey=0A>and endkey. For example, if these are [201=
2, 4, 5] and [2014, 8, 3]=0A>respectively, we want to include [2013, 3, 6] =
in the output even=0A>though its=0A>second element does not lie between the=
 second elements of the keys.=0A>=C2=A0=0A>Nick=0A>=C2=A0=0A>=C2=A0=0A>On 2=
 April 2014 03:56, Schroiff, Klaus <KlausS@fast.au.fujitsu.com>=0A>wrote:=
=0A>=C2=A0=0A>Hi,=0A>=C2=A0=0A>Let's assume that I have the following view =
function:=0A>=C2=A0=0A>function(doc) { emit([a, b], whatever) } }=0A>=C2=A0=
=0A>In the query I'm running something like=0A>startkey=3D[a1, b1] & endkey=
=3D[a2,b2]=0A>Thus there're TWO explicit ranges here - a1->a2 and b1->b2.=
=0A>=C2=A0=0A>What is the expected result ?=0A>=C2=A0=0A>My understanding:=
=0A>The query is executed in two phases=0A>In the first phase, the index is=
 filtered for qualifying results=0A>where the=0A>first key ranges from a1 t=
o a2=0A>In the second phase, these filtered results are filtered once more=
=0A>according to the range of the second emitted key from b1 to b2.=0A>Thus=
 essentially an "AND" operation.=0A>The filtering is performed using lexico=
graphical rules.=0A>=C2=A0=0A>Is that correct ? The doc about complex keys =
is a bit slim.=0A>=C2=A0=0A>Thanks=0A>=C2=A0=0A>Klaus=0A>=C2=A0=0A>=C2=A0
---1672125993-556949270-1396529627=:10521--