Return-Path: Delivered-To: 2-oracle-l@orafaq.com Received: (qmail 15266 invoked from network); 20 Sep 2007 13:23:16 -0500 Received: from freelists-180.iquest.net (HELO turing.freelists.org) (206.53.239.180) by 69.64.49.119 with SMTP; 20 Sep 2007 13:23:11 -0500 Received: from localhost (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id B59F2760D5A; Thu, 20 Sep 2007 14:23:02 -0400 (EDT) Received: from turing.freelists.org ([127.0.0.1]) by localhost (turing.freelists.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 06635-01; Thu, 20 Sep 2007 14:23:02 -0400 (EDT) Received: from turing (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 19221760C22; Thu, 20 Sep 2007 14:23:02 -0400 (EDT) Received: with ECARTIS (v1.0.0; list oracle-l); Thu, 20 Sep 2007 13:37:33 -0400 (EDT) Received: from localhost (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 70ECE760FAD for ; Thu, 20 Sep 2007 13:37:33 -0400 (EDT) Received: from turing.freelists.org ([127.0.0.1]) by localhost (turing.freelists.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 30843-06 for ; Thu, 20 Sep 2007 13:37:33 -0400 (EDT) Received: from nz-out-0506.google.com (nz-out-0506.google.com [64.233.162.226]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 251D6760FC6 for ; Thu, 20 Sep 2007 13:37:32 -0400 (EDT) Received: by nz-out-0506.google.com with SMTP id s18so495409nze for ; Thu, 20 Sep 2007 10:37:32 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=beta; h=domainkey-signature:received:received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; bh=qxArp3+NdfAgJYWwCtg6gWxh7DnCvp2ZB4UZIrEYJiA=; b=BW+RpK4qxyedmPkmaCbiIpSTHE7yPqzZuBacEt7SHZy7Hj88D2FSlxRUPQIwtyhCiSbCi9SBRSIo6lAlvfYgOuN6zzTTzlOK2P/a1hHOYnVVjXUBppzPNGfwuwgf8RjqEfcCalr0LmsrVq3aBJq5JQln+NusRHbQyqcDYCuo5Ko= DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=beta; h=received:message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:references; b=SsDa+UC48aKoScN+jvgbtufa6xZ1xYh+VUUc7ko+HM8EyQTUIGJx5rikpTBjiRyt13LjB03/WF2H2st70CEH3n4tHr+9XUXm2MzNHq611dTv1zfJlbdC9OOgayQZZT+yIOgGov620qqwS65hxDK1kSgvJGBfAOnfHZK+uE8gxIw= Received: by 10.142.231.7 with SMTP id d7mr922085wfh.1190309851749; Thu, 20 Sep 2007 10:37:31 -0700 (PDT) Received: by 10.143.39.20 with HTTP; Thu, 20 Sep 2007 10:37:31 -0700 (PDT) Message-ID: <9c9b9dc90709201037o18a47b8dt58a85b1746fcd5f7@mail.gmail.com> Date: Thu, 20 Sep 2007 13:37:31 -0400 From: "Rumpi Gravenstein" To: nigel_cl_thomas@yahoo.com Subject: Re: am I reading this SQL correctly? Cc: genegurevich@discover.com, oracle-l In-Reply-To: <888447.42589.qm@web58712.mail.re1.yahoo.com> MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_Part_27153_23420195.1190309851734" References: <888447.42589.qm@web58712.mail.re1.yahoo.com> X-archive-position: 1770 X-ecartis-version: Ecartis v1.0.0 Sender: oracle-l-bounce@freelists.org Errors-to: oracle-l-bounce@freelists.org X-original-sender: rgravens@gmail.com Precedence: normal Reply-to: rgravens@gmail.com List-help: List-unsubscribe: List-software: Ecartis version 1.0.0 List-Id: oracle-l X-List-ID: oracle-l List-subscribe: List-owner: List-post: List-archive: X-list: oracle-l X-Virus-Scanned: Debian amavisd-new at localhost.localdomain ------=_Part_27153_23420195.1190309851734 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline One issue that Nigel hasn't addressed is that you will want to account for a call overlapping with itself. That might be done with something along the lines of: a.ROWID != b.ROWID On 9/20/07, Nigel Thomas wrote: > > Gene > > 1) How is the "current call" identified? - this query will report ANY > calls that overlap (where call A starts during call B) - even if both calls > are way in the past. > 2) It could be restated as a.CALL_START_TS between B.CALL_START_TS and > B.CALL_END_TS - which could be a range scan (in a nested loop) but B has > no predicates. Still unpleasant! > 3) Why doesn't it check for B starting during A > 4) Why doesn't it check for B completely overlapping A (B starts before > and finishes after A) and vice versa > > You have identified: > A: ---- | -- | ---- | ---- > B:---- | ------ | ---- | ---- > > but not other variants eg > > A:---- | ------ | ---- > B: ---- | --- | ---- > > (view in a monotype font or this will look silly) > > 5) In a procedural loop (within a pipeline function) I think you could do > this in a single pass (ordered by CALL_START_TS) > start with an empty in memory collection of calls (start time, end time, > overlap count) > for all records > - read a record > - increment overlap count for all collected calls that overlap with this > one > - register its start time and end time (into the collection in memory) > (with the total overlap count just calculated) > - report (eg by emitting a row from the pipelined function) all calls in > the collection with an end time prior to the current record's start (don't > include records with no overlaps) > - and remove those records from the register > At the end of the loop, emit all remaining records with overlap count > 0 > (possibly grouping by call_start_ts - as several calls could start > simultaneously I presume) > > Having described that, is it possible to do the same with a analytic > functions (eg WINDOW)? > > HTH > > Regards Nigel > > ----- Original Message ---- > From: "genegurevich@discover.com" genegurevich@discover.com > > I am trying to tune the following SQL (in 10.2.0.3) > > SELECT a.CALL_STRT_TS, count(b.Call_STRT_TS) > from IVR.IVC_IVR_CMS_CALL a, IVR.IVC_IVR_CMS_CALL > where a.CALL_STRT_TS >= b.CALL_STRT_TS > and a.CALL_STRT_TS <= b.CALL_END_TS > group by a.CALL_STRT_TS; > > The goal of this is to count the number of calls that overlap timewise > with a current call. > -- > http://www.freelists.org/webpage/oracle-l > > > -- Rumpi Gravenstein ------=_Part_27153_23420195.1190309851734 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline
One issue that Nigel hasn't addressed is that you will want to account for a call overlapping with itself.  That might be done with something along the lines of:
 
a.ROWID != b.ROWID
 
On 9/20/07, Nigel Thomas <nigel_cl_thomas@yahoo.com> wrote:
Gene

1) How is the "current call" identified? - this query will report ANY calls that overlap (where call A starts during call B) - even if both calls are way in the past.
2) It could be restated as a.CALL_START_TS between B.CALL_START_TS and B.CALL_END_TS - which could be a range scan (in a nested loop)  but B has no predicates. Still unpleasant!
3) Why doesn't it check for B starting during A
4) Why doesn't it check for B completely overlapping A (B starts before and finishes after A) and vice versa

You have identified:
A:  ----  |   --    |  ----  |     ----
B:----    | ------  |  ----  | ----

but not other variants eg

A:----    | ------  | ----
B:  ----  |   ---   |     ----

(view in a monotype font or this will look silly)

5) In a procedural loop (within a pipeline function) I think you could do this in a single pass (ordered by CALL_START_TS)
start with an empty in memory collection of calls (start time, end time, overlap count)
for all records
- read a record
- increment overlap count for all collected calls that overlap with this one
- register its start time and end time (into the collection in memory) (with the total overlap count just calculated)
- report (eg by emitting a row from the pipelined function) all calls in the collection with an end time prior to the current record's start (don't include records with no overlaps)
- and remove those records from the register
At the end of the loop, emit all remaining records with overlap count > 0 (possibly grouping by call_start_ts - as several calls could start simultaneously I presume)

Having described that, is it possible to do the same with a analytic functions (eg WINDOW)?

HTH

Regards Nigel

----- Original Message ----
From: "genegurevich@discover.com" genegurevich@discover.com

I am trying to tune the following SQL (in 10.2.0.3)

SELECT a.CALL_STRT_TS, count(b.Call_STRT_TS)
from IVR.IVC_IVR_CMS_CALL  a, IVR.IVC_IVR_CMS_CALL
where a.CALL_STRT_TS >= b.CALL_STRT_TS
   and a.CALL_STRT_TS <= b.CALL_END_TS
group by a.CALL_STRT_TS;

The goal of this is  to count the number of calls that overlap timewise
with a current call.
--
http://www.freelists.org/webpage/oracle-l





--
Rumpi Gravenstein ------=_Part_27153_23420195.1190309851734-- -- http://www.freelists.org/webpage/oracle-l