Return-Path: X-Original-To: apmail-hive-user-archive@www.apache.org Delivered-To: apmail-hive-user-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 0C56417D39 for ; Tue, 17 Mar 2015 15:55:14 +0000 (UTC) Received: (qmail 40517 invoked by uid 500); 17 Mar 2015 15:55:09 -0000 Delivered-To: apmail-hive-user-archive@hive.apache.org Received: (qmail 40451 invoked by uid 500); 17 Mar 2015 15:55:09 -0000 Mailing-List: contact user-help@hive.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: user@hive.apache.org Delivered-To: mailing list user@hive.apache.org Received: (qmail 40441 invoked by uid 99); 17 Mar 2015 15:55:09 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 17 Mar 2015 15:55:09 +0000 X-ASF-Spam-Status: No, hits=3.2 required=5.0 tests=FORGED_YAHOO_RCVD,HTML_MESSAGE,RCVD_IN_DNSWL_NONE,SPF_PASS X-Spam-Check-By: apache.org Received-SPF: pass (nike.apache.org: domain of ququrico@yahoo.com designates 98.136.217.113 as permitted sender) Received: from [98.136.217.113] (HELO nm25-vm2.bullet.mail.gq1.yahoo.com) (98.136.217.113) by apache.org (qpsmtpd/0.29) with ESMTP; Tue, 17 Mar 2015 15:54:42 +0000 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=yahoo.com; s=s2048; t=1426607490; bh=YeNkP1uySlgjNy/bod1nD7DgPBR+qBM8rX7NRarNimY=; h=Date:From:Reply-To:To:In-Reply-To:References:Subject:From:Subject; b=FAHBOEXMzWIf7CoyucRi6owIfB5qD5IDtOSXfs0mDaC1oDLgOwEisoUckdU88J/Je+kk7BKY791smlVkWak2PpQi5gVzC/LLcYv/dwSeW3OO3r/rIfNsZ2FPCp/qyxmIsTzvtBPMtHZUBHz/AjsOV+xehGKDlIeQU5baClyY6UDEqD9eU86WVFLMxoMwKGjPG4Wgr0n2YykSbb/hot2w2r+RgFaEz3DkUVyRXKLJvco9qEj9LciNBQYlbYMAu/Yr5ENF3uOAQ1geqcc525kF3N3BhsA6ixSeohuugQuv9Ro0sei5sma3t5dG1AGUv9xLpHJAaNSxuQWwDBrHY9u6Vg== Received: from [98.137.12.62] by nm25.bullet.mail.gq1.yahoo.com with NNFMP; 17 Mar 2015 15:51:30 -0000 Received: from [98.137.12.228] by tm7.bullet.mail.gq1.yahoo.com with NNFMP; 17 Mar 2015 15:51:30 -0000 Received: from [127.0.0.1] by omp1036.mail.gq1.yahoo.com with NNFMP; 17 Mar 2015 15:51:30 -0000 X-Yahoo-Newman-Property: ymail-3 X-Yahoo-Newman-Id: 10621.206.bm@omp1036.mail.gq1.yahoo.com X-YMail-OSG: ZmezHMcVM1kLqdmd3QO4QnqUCrL07uOtrp2O8q788c9n30lLQGzP2wEmwkKLVHZ 6FIlFdzWz4mDlY.6jI6V19t9lwjEczGQ_wfZDmirvkaZ0ZW41VgbijDRRWo4P7rLe8Zdd4NoCFBU COQxcUhGaMpLWyyjLzb6qX1iqzRJeNrpI7DS_MKFyYEscxlpTw5GyZcNZod8XJaB0MpHThGtT.lL TQ.2aRJhOEG51U9tJuJHhdm9WADa1llBeCJsF.5nG5DRcbeFa.URbcJNdJMMEGRoF1x61dXHUFkW CSqIbOMftsjNgeD3qh2lZ_OdNMgH0S4xhnzVszE4XaR2gyOWJcjvV6XsV88FlwhiqauyxsTFIDh0 C9tB6FysejlVZNii_Q15gccRdqS.mQ48EkxqbSUWyZmqgkeHE.MWGxu_qKdH6XSdZ_g3Q.MYYnFp kV2Beb8NOyrhEKzFEcYvWvUjibUXmfmtQHf7EGL0yJI0my595UodqvHcAwzRpSrI6Lr6eTxRuGiB udw-- Received: by 98.137.12.52; Tue, 17 Mar 2015 15:51:29 +0000 Date: Tue, 17 Mar 2015 15:51:29 +0000 (UTC) From: cobby cohen Reply-To: cobby cohen To: "user@hive.apache.org" Message-ID: <285622750.133338.1426607489094.JavaMail.yahoo@mail.yahoo.com> In-Reply-To: <03a701d06020$7ea336f0$7be9a4d0$@co.uk> References: <03a701d06020$7ea336f0$7be9a4d0$@co.uk> Subject: Re: view over partitioned table MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_Part_133337_889066787.1426607489063" X-Virus-Checked: Checked by ClamAV on apache.org ------=_Part_133337_889066787.1426607489063 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable we did empirical test on a partitioned table with ~50 million rows, and fil= tering from a view (which includes an explode() on an array column) vs filt= ering the underlining table on a specific partition yielding similar result= s.i also noticed that an "explain dependency" statement on both the view an= d the table shows the analizer does indeed pick up the right partition on b= oth cases (view and table). From: Mich Talebzadeh To: user@hive.apache.org; 'cobby cohen' =20 Sent: Monday, March 16, 2015 9:36 PM Subject: RE: view over partitioned table =20 #yiv3664821214 #yiv3664821214 -- _filtered #yiv3664821214 {font-family:Helv= etica;panose-1:2 11 6 4 2 2 2 2 2 4;} _filtered #yiv3664821214 {panose-1:2 = 4 5 3 5 4 6 3 2 4;} _filtered #yiv3664821214 {font-family:Calibri;panose-1:= 2 15 5 2 2 2 4 3 2 4;} _filtered #yiv3664821214 {font-family:Tahoma;panose-= 1:2 11 6 4 3 5 4 4 2 4;}#yiv3664821214 #yiv3664821214 p.yiv3664821214MsoNor= mal, #yiv3664821214 li.yiv3664821214MsoNormal, #yiv3664821214 div.yiv366482= 1214MsoNormal {margin:0cm;margin-bottom:.0001pt;font-size:12.0pt;}#yiv36648= 21214 a:link, #yiv3664821214 span.yiv3664821214MsoHyperlink {color:blue;tex= t-decoration:underline;}#yiv3664821214 a:visited, #yiv3664821214 span.yiv36= 64821214MsoHyperlinkFollowed {color:purple;text-decoration:underline;}#yiv3= 664821214 p.yiv3664821214MsoAcetate, #yiv3664821214 li.yiv3664821214MsoAcet= ate, #yiv3664821214 div.yiv3664821214MsoAcetate {margin:0cm;margin-bottom:.= 0001pt;font-size:8.0pt;}#yiv3664821214 p.yiv3664821214msonormal, #yiv366482= 1214 li.yiv3664821214msonormal, #yiv3664821214 div.yiv3664821214msonormal {= margin-right:0cm;margin-left:0cm;font-size:12.0pt;}#yiv3664821214 p.yiv3664= 821214msochpdefault, #yiv3664821214 li.yiv3664821214msochpdefault, #yiv3664= 821214 div.yiv3664821214msochpdefault {margin-right:0cm;margin-left:0cm;fon= t-size:12.0pt;}#yiv3664821214 span.yiv3664821214msohyperlink {}#yiv36648212= 14 span.yiv3664821214msohyperlinkfollowed {}#yiv3664821214 span.yiv36648212= 14emailstyle17 {}#yiv3664821214 p.yiv3664821214msonormal1, #yiv3664821214 l= i.yiv3664821214msonormal1, #yiv3664821214 div.yiv3664821214msonormal1 {marg= in:0cm;margin-bottom:.0001pt;font-size:12.0pt;}#yiv3664821214 span.yiv36648= 21214msohyperlink1 {color:blue;text-decoration:underline;}#yiv3664821214 sp= an.yiv3664821214msohyperlinkfollowed1 {color:purple;text-decoration:underli= ne;}#yiv3664821214 span.yiv3664821214emailstyle171 {color:windowtext;}#yiv3= 664821214 p.yiv3664821214msochpdefault1, #yiv3664821214 li.yiv3664821214mso= chpdefault1, #yiv3664821214 div.yiv3664821214msochpdefault1 {margin-right:0= cm;margin-left:0cm;font-size:10.0pt;}#yiv3664821214 span.yiv3664821214Ballo= onTextChar {}#yiv3664821214 span.yiv3664821214EmailStyle29 {color:windowtex= t;}#yiv3664821214 .yiv3664821214MsoChpDefault {font-size:10.0pt;} _filtered= #yiv3664821214 {margin:72.0pt 72.0pt 72.0pt 72.0pt;}#yiv3664821214 div.yiv= 3664821214WordSection1 {}#yiv3664821214 Hi Cobby, =C2=A0In your statement a= nd I quote: =C2=A0=E2=80=9CEmpirical tests, by the show, suggest that indee= d, partition optimization does take place. =E2=80=9C =C2=A0Do you have an e= xample of this please? =C2=A0Thanks, =C2=A0Mich =C2=A0Creating in-memory D= ata Grid for Trading Systems with Oracle TimesTen and Coherence Cache =C2= =A0NOTE: The information in this email is proprietary and confidential. Thi= s message is for the designated recipient only, if you are not the intended= recipient, you should destroy it immediately. Any information in this mess= age shall not be understood as given or endorsed by Peridale Ltd, its subsi= diaries or their employees, unless expressly so stated. It is the responsib= ility of the recipient to ensure that this email is virus free, therefore n= either Peridale Ltd, its subsidiaries nor their employees accept any respon= sibility. =C2=A0 From: cobby cohen [mailto:ququrico@yahoo.com]=20 Sent: 16 March 2015 16:14 To: user@hive.apache.org Subject: Re: view over partitioned table =C2=A0hi,i am using CDH 5.2 (hive = 13).as for partitioned views, to my understanding they are more about abstr= acting a non partitioned back end. My questing is related to how and if the= query optimizer can detect possible uses of partitioned column, even if it= might be "masked" by a view (or multiple view).Empirical tests, by the sho= w, suggest that indeed, partition optimization does take place. =C2=A0regar= ds, cobby. =C2=A0From: "Moore, Douglas" To: "user@hive.apache.org" ; 'cobby cohen' =20 Sent: Monday, March 16, 2015 4:19 PM Subject: Re: view over partitioned table =C2=A0Mich, =C2=A0What version of = Hive are you running?Have you seen this?https://cwiki.apache.org/confluence= /display/Hive/PartitionedViews =C2=A0- Douglas =C2=A0From: Mich Talebzadeh = Reply-To: Date: Sun, 15 Mar 2015 19:01:57 +0000 To: , 'cobby cohen' Subject: RE: view over partitioned table =C2=A0OK,=C2=A0This is the way I r= ead it. Crearte table t1 .. partitioned by date will use horizontal partiti= oning as per common with any RDBMS say Oracle.=C2=A0The view I will create = it as follows:=C2=A0hive> create view v1 as select * from t1;OKTime taken: = 0.073 seconds=C2=A0hive> analyze table t1 partition (date) compute statisti= cs;Query ID =3D hduser_20150315184646_af3890b2-4079-4f81-b38b-da9af67b2751T= otal jobs =3D 1Launching Job 1 out of 1-----Stage-Stage-0: Map: 1=C2=A0=C2= =A0 Cumulative CPU: 0.6 sec=C2=A0=C2=A0 HDFS Read: 299 HDFS Write: 0 SUCCES= STotal MapReduce CPU Time Spent: 600 msecOK=C2=A0hive> insert into table t1= partition (date)=C2=A0 values (1,'2015-04-04');hive> insert into table t1 = partition (date)=C2=A0 values (1,'2015-05-04');=C2=A0Help to update stats= =C2=A0hive> analyze table t1 partition (date) compute statistics;=C2=A0sele= ct from the view after exiting hive and back=C2=A0=C2=A0hive> explain selec= t id from v1 where date =3D '2015-04-04';OKSTAGE DEPENDENCIES:=C2=A0 Stage-= 0 is a root stage=C2=A0STAGE PLANS:=C2=A0 Stage: Stage-0=C2=A0=C2=A0=C2=A0 = Fetch Operator=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 limit: -1=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 Processor Tree:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 TableSc= an=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 alias: t1=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Statistics: Num rows: 1 Data = size: 1 Basic stats: COMPLETE Column stats: NONE=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 Select Operator=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 expressions: id (type: string)=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 outputColumnNa= mes: _col0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0 Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column stats= : NONE=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Li= stSink=C2=A0Time taken: 0.627 seconds, Fetched: 17 row(s)=C2=A0=C2=A0Now se= lect the same from base table=C2=A0hive> explain select id from t1 where da= te =3D '2015-04-04';OKSTAGE DEPENDENCIES:=C2=A0 Stage-0 is a root stage=C2= =A0STAGE PLANS:=C2=A0 Stage: Stage-0=C2=A0=C2=A0=C2=A0 Fetch Operator=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 limit: -1=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Processor = Tree:=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 TableScan=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 alias: t1=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0 Statistics: Num rows: 1 Data size: 1 Basic stat= s: COMPLETE Column stats: NONE=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0 Select Operator=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0 expressions: id (type: string)=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 outputColumnNames: _col0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 Statistics: Nu= m rows: 1 Data size: 1 Basic stats: COMPLETE Column stats: NONE=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 ListSink=C2=A0Time t= aken: 0.597 seconds, Fetched: 17 row(s)=C2=A0The plan looks the same but it= is slower on the view=C2=A0HTH=C2=A0Mich =C2=A0NOTE: The information in th= is email is proprietary and confidential. This message is for the designate= d recipient only, if you are not the intended recipient, you should destroy= it immediately. Any information in this message shall not be understood as= given or endorsed by Peridale Ltd, its subsidiaries or their employees, un= less expressly so stated. It is the responsibility of the recipient to ensu= re that this email is virus free, therefore neither Peridale Ltd, its subsi= diaries nor their employees accept any responsibility.=C2=A0From: cobby coh= en [mailto:ququrico@yahoo.com]=20 Sent: 15 March 2015 17:52 To: user@hive.apache.org Subject: view over partitioned table=C2=A0in the following case:=C2=A0creat= e table T1(id string)partitioned by (date date);=C2=A0create view V1(id, da= te)select id, date from T1=C2=A0if i do the followingselect id from V1 wher= e date =3D '2015-04-04'will the query optimize on T1's partitioning?=C2=A0a= nd what if V1 is a bit more complex, further depending on other views (all = of which expose the partitioned column as-is)?=C2=A0thx! =C2=A0 ------=_Part_133337_889066787.1426607489063 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
we = did empirical test on a partitioned table with ~50 million rows, and filter= ing from a view (which includes an explode() on an array column) vs filteri= ng the underlining table on a specific partition yielding similar results.<= /div>
i also notic= ed that an "explain dependency" statement on both the view and the table sh= ows the analizer does indeed pick up the right partition on both cases (vie= w and table).

From: Mi= ch Talebzadeh <mich@peridale.co.uk>
To: user@hive.apache.org; 'cobby cohen' <ququrico@ya= hoo.com>
Sent: Mon= day, March 16, 2015 9:36 PM
Subje= ct: RE: view over partitioned table

Hi Cobby,
 
In your statement and I quote:
 
=E2=80=9CEmpirical tests, by the show, suggest that indeed, p= artition optimization does take place. =E2=80=9C
 
Do you have an example of t= his please?
 
Thanks,
 
Mich
 <= /span>
Creating in-me= mory Data Grid for Trading Systems with Oracle TimesTen and Coherence Cache=
 
NOTE: The information in this email is pro= prietary and confidential. This message is for the designated recipient onl= y, if you are not the intended recipient, you should destroy it immediately= . Any information in this message shall not be understood as given or endor= sed by Peridale Ltd, its subsidiaries or their employees, unless expressly = so stated. It is the responsibility of the recipient to ensure that this em= ail is virus free, therefore neither Peridale Ltd, its subsidiaries nor the= ir employees accept any responsibility.
 


From: cobby cohen [mailto:ququrico@yahoo.com]
Sent: 16 March 2015 16:14
To: user@hive.apache.org
Subject: Re: view over partitioned tab= le
 
hi,
i am usi= ng CDH 5.2 (hive 13).
as for partitioned views, to my unde= rstanding they are more about abstracting a non partitioned back end. My qu= esting is related to how and if the query optimizer can detect possible use= s of partitioned column, even if it might be "masked" by a view (or multipl= e view).
Empirical tests, by the show, suggest that indeed,= partition optimization does take place.  
regards, = cobby.
=  
From: "Moore, Douglas" <<= a rel=3D"nofollow" shape=3D"rect" ymailto=3D"mailto:Douglas.Moore@thinkbiga= nalytics.com" target=3D"_blank" href=3D"mailto:Douglas.Moore@thinkbiganalyt= ics.com">Douglas.Moore@thinkbiganalytics.com>
To: "user@hive.apache.org" <user@hive.apache.org= >; 'cobby cohen' <= ququrico@yahoo.com>
Sent: Monday, March 16= , 2015 4:19 PM
Subject: Re: view over partitioned = table
 
Mich,
=  
= What version of Hive are you running?
Have you seen this?
 
- Douglas
 
<= div id=3D"yiv3664821214yqt17287">
From: Mich Talebzadeh <mich@peridale.co.uk>
Rep= ly-To: <user@= hive.apache.org>
Date: Sun, 15 Mar 2015 19:= 01:57 +0000
To: <user@hive.apache.org>, 'cobby cohen' <ququrico@yahoo.com>
= Subject: RE: view over partitioned table
 
<= /div>
=
OK,
 
This is t= he way I read it. Crearte table t1 .. partitioned by date will use horizont= al partitioning as per common with any RDBMS say Oracle.
 
The view I will create it as follows= :
&= nbsp;
hive> create vi= ew v1 as select * from t1;
<= div class=3D"yiv3664821214MsoNormal" style=3D"background:white;">OK
Time taken: 0.073 seconds
 
hive> analyze table t1 partit= ion (date) compute statistics;
Query ID =3D hduser_20150315184646_af3890b2-407= 9-4f81-b38b-da9af67b2751
Total jobs =3D 1
Launching Job 1 out of 1
-----
Stage-Stage-0: Map: 1   C= umulative CPU: 0.6 sec   HDFS Read: 299 HDFS Write: 0 SUCCESS
Total = MapReduce CPU Time Spent: 600 msec
OK
 
=
hive> insert into table t1 partition= (date)  values (1,'2015-04-04');
=
hive> insert into table t1 partition= (date)  values (1,'2015-05-04');
=
 
Help to update stats
 
hive> anal= yze table t1 partition (date) compute statistics;
 
select from the view aft= er exiting hive and back
 
 
hive> explain select id from v1 where dat= e =3D '2015-04-04';
OK
STAGE DEPENDENCIES:
=   Stage-0 is a root stage
 =
STAGE PLA= NS:
  Stage: Stage-0
    Fetch Operator
    =   limit: -1
      Processor Tree:
  &nbs= p;     TableScan
=        = ;   alias: t1
        &nbs= p; Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column stats:= NONE
          Select Operator
&nbs= p;           expressions:= id (type: string)
          = ;  outputColumnNames: _col0
=
       &nb= sp;    Statistics: Num rows: 1 Data size: 1 Basic stats: COM= PLETE Column stats: NONE
        &nbs= p;   ListSink
 
Time taken: 0.627 seconds, Fetched: 17 row(s)
 = ;
&= nbsp;
Now select the same from base table
=
 
hive> explain select id from t1 = where date =3D '2015-04-04';
OK
STAGE DEPENDENCIES:
  Stage-0 is a root stage
&nbs= p;
= STAGE PLANS:
  Stage: Stage-0
<= div>
    Fetch Operator
  &nb= sp;   limit: -1
      Processor Tree:
 &n= bsp;      TableScan=
     &nb= sp;    alias: t1
<= div>
       &nbs= p;  Statistics: Num rows: 1 Data size: 1 Basic stats: COMPLETE Column = stats: NONE
          Select Oper= ator
            expres= sions: id (type: string)
        &nbs= p;   outputColumnNames: _col0
      &n= bsp;     Statistics: Num rows: 1 Data size: 1 Basic sta= ts: COMPLETE Column stats: NONE
<= div>
       &nbs= p;    ListSink
 
Time taken: 0.597 seconds, Fetched: 17 row(s= )
&= nbsp;
The plan looks the same but it is slower on the view
 
HTH
 =
Mich
 
NOTE: The information in this email is proprietary and confidential. T= his message is for the designated recipient only, if you are not the intend= ed recipient, you should destroy it immediately. Any information in this me= ssage shall not be understood as given or endorsed by Peridale Ltd, its sub= sidiaries or their employees, unless expressly so stated. It is the respons= ibility of the recipient to ensure that this email is virus free, therefore= neither Peridale Ltd, its subsidiaries nor their employees accept any resp= onsibility.
 
= From: cobby cohen [mailto:ququrico@yahoo.com]
Sent: 15 March 2015 17:52
To: user@hive.apache.orgSubject: view over partitioned table
 
in the following case:
 
cr= eate table T1(id string)
partitioned by (date date);
<= span style=3D""> 
create view V1(id, date)
select id, date from T1
 
if i do the following
select id from V1 where date= =3D '2015-04-04'
will the query optimize on T1's partit= ioning?
 
and what if V1 is a bi= t more complex, further depending on other views (all of which expose the p= artitioned column as-is)?
 
th= x!
 
=


------=_Part_133337_889066787.1426607489063--