From oracle-l-bounce@freelists.org Fri Sep 9 13:34:07 2005 Return-Path: Received: from air891.startdedicated.com (root@localhost) by orafaq.com (8.12.10/8.12.10) with ESMTP id j89IY79N023186 for ; Fri, 9 Sep 2005 13:34:07 -0500 X-ClientAddr: 206.53.239.180 Received: from turing.freelists.org (freelists-180.iquest.net [206.53.239.180]) by air891.startdedicated.com (8.12.10/8.12.10) with ESMTP id j89IXfIP023121 for ; Fri, 9 Sep 2005 13:33:47 -0500 Received: from localhost (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id 6A88B1EB667; Fri, 9 Sep 2005 13:31:40 -0500 (EST) Received: from turing.freelists.org ([127.0.0.1]) by localhost (turing [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id 18402-05; Fri, 9 Sep 2005 13:31:40 -0500 (EST) Received: from turing (localhost [127.0.0.1]) by turing.freelists.org (Avenir Technologies Mail Multiplex) with ESMTP id D836F1EB666; Fri, 9 Sep 2005 13:31:39 -0500 (EST) Message-ID: <5B257A26B4845C469B87871B6CEFE5070331AD59@usnjc04wmx003.tdwaterhouse.com> From: JayMiller@TDWaterhouse.com To: DGoulet@vicr.com, oracle-l@freelists.org Subject: RE: Query started getting 1722 error - interesting explanation wh y Date: Fri, 9 Sep 2005 14:29:16 -0400 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----_=_NextPart_001_01C5B56C.62D1DEB0" X-archive-position: 25241 X-ecartis-version: Ecartis v1.0.0 Sender: oracle-l-bounce@freelists.org Errors-To: oracle-l-bounce@freelists.org X-original-sender: JayMiller@TDWaterhouse.com Precedence: normal Reply-To: JayMiller@TDWaterhouse.com X-list: oracle-l X-Virus-Scanned: by amavisd-new-20030616-p9 (Debian) at avenirtech.net X-mailscan-MailScanner-Information: Please contact the ISP for more information X-mailscan-MailScanner: Found to be clean X-MailScanner-From: oracle-l-bounce@freelists.org X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.63 (2004-01-11) on air891.startdedicated.com X-Spam-Status: No, hits=-3.1 required=5.0 tests=AWL,BAYES_00,HTML_50_60, HTML_FONTCOLOR_UNKNOWN,HTML_MESSAGE,NO_REAL_NAME autolearn=no version=2.63 ------_=_NextPart_001_01C5B56C.62D1DEB0 Content-Type: text/plain Hmm, I think I checked that at the time but a week and a half later I can't say for certain. Currently that is not the case but currently it's back to the old explain plan as well so that doesn't say anything. Thanks, Jay Miller Sr. Oracle DBA x68355 -----Original Message----- From: Goulet, Dick [mailto:DGoulet@vicr.com] Sent: Friday, September 09, 2005 2:23 PM To: Miller, Jay; oracle-l@freelists.org Subject: RE: Query started getting 1722 error - interesting explanation why Jay, Did that varchar2 have numbers only in it before it stopped working and then picked up one with an alpha character? I've seen that one more than I care to remember. _____ From: oracle-l-bounce@freelists.org [mailto:oracle-l-bounce@freelists.org] On Behalf Of JayMiller@TDWaterhouse.com Sent: Friday, September 09, 2005 2:16 PM To: oracle-l@freelists.org Subject: Query started getting 1722 error - interesting explanation why I got a call from a developer the other week complaining that a query in QA had started giving an ora-1722. It had worked fine for several months and the identical query was working fine in production. Select count(1) from prod.job_profile jp where step_id < 9999 and not exists (select 'x' from prod.job_log jl where jl.job_id = jp.job_id and ( jl.status in ('C', 'EI') ) and jl.batch_id in (select to_number(text_value) from prod.batch_parameter where name = 'BATCH_ID')); Investigating confirmed this, along with the fact that the query was comparing a string (jl.batch_id) with a number (to_number(text_value)). But why did it suddenly stop working and why did it still work in production? Comparing explain plans between QA and production I saw that they were different (statistics had changed during QA testing). The query that failed was using an index on jl.batch_id. Apparently Oracle automatically took care of converting the string to a number *unless* the number was being used to access an index on the varchar2 column. Interesting behavior. I can see why it might work that way but it's certainly not intuitive. Of course explaining this to the developer took quite a while. "It worked before, why shouldn't it work now? What's wrong with Oracle?" Oracle 9.2.0.6 Jay Miller Sr. Oracle DBA . _____ This message is confidential and sent by TD Waterhouse solely for use by the intended recipient. If you are not the intended recipient, you are hereby notified that any use, distribution or copying of this communication is strictly prohibited. This should not be deemed as an offer or solicitation, to buy or sell any product. Any 3rd party information contained herein was prepared by sources deemed reliable, but is not guaranteed. TD Waterhouse does not accept electronic instructions that would require an original signature. Information received by or sent from TD Waterhouse is stored, subject to review, and may be produced to regulatory authorities or others with a legal right to such. ------_=_NextPart_001_01C5B56C.62D1DEB0 Content-Type: text/html

Hmm, I think I checked that at the time but a week and a half later I can't say for certain.  Currently that is not the case but currently it's back to the old explain plan as well so that doesn't say anything.

 

Thanks,

Jay Miller

Sr. Oracle DBA

x68355

 

-----Original Message-----
From: Goulet, Dick [mailto:DGoulet@vicr.com]
Sent: Friday, September 09, 2005 2:23 PM
To: Miller, Jay; oracle-l@freelists.org
Subject: RE: Query started getting 1722 error - interesting explanation why

 

Jay,

 

    Did that varchar2 have numbers only in it before it stopped working and then picked up one with an alpha character?  I've seen that one more than I care to remember.

 

From: oracle-l-bounce@freelists.org [mailto:oracle-l-bounce@freelists.org] On Behalf Of JayMiller@TDWaterhouse.com
Sent: Friday, September 09, 2005 2:16 PM
To: oracle-l@freelists.org
Subject: Query started getting 1722 error - interesting explanation why

I got a call from a developer the other week complaining that a query in QA had started giving an ora-1722.  It had worked fine for several months and the identical query was working fine in production.

 

Select count(1)

from prod.job_profile jp

where step_id < 9999

and not exists (select 'x' from prod.job_log jl

                where jl.job_id = jp.job_id

                and ( jl.status in ('C', 'EI') )

                and  jl.batch_id in (select to_number(text_value)

                                     from prod.batch_parameter

                                     where name = 'BATCH_ID'));

 

Investigating confirmed this, along with the fact that the query was comparing a string (jl.batch_id) with a number (to_number(text_value)).  But why did it suddenly stop working and why did it still work in production?

 

Comparing explain plans between QA and production I saw that they were different (statistics had changed during QA testing).  The query that failed was using an index on jl.batch_id. 

 

Apparently Oracle automatically took care of converting the string to a number *unless* the number was being used to access an index on the varchar2 column.  Interesting behavior.  I can see why it might work that way but it's certainly not intuitive.

 

Of course explaining this to the developer took quite a while.  "It worked before, why shouldn't it work now?  What's wrong with Oracle?"

 

Oracle 9.2.0.6

 

 

Jay Miller

Sr. Oracle DBA

 

.

This message is confidential and sent by TD Waterhouse solely for use
by the intended recipient. If you are not the intended recipient, you
are hereby notified that any use, distribution or copying of this
communication is strictly prohibited. This should not be deemed as an
offer or solicitation, to buy or sell any product. Any 3rd party
information contained herein was prepared by sources deemed reliable,
but is not guaranteed. TD Waterhouse does not accept electronic
instructions that would require an original signature. Information
received by or sent from TD Waterhouse is stored, subject to review,
and may be produced to regulatory authorities or others with a legal
right to such.

------_=_NextPart_001_01C5B56C.62D1DEB0-- -- http://www.freelists.org/webpage/oracle-l