Comments for Oracloid Blog http://www.oracloid.com The Alex Gorbachev Oracle Blog Sat, 11 Sep 2010 02:12:18 +0000 http://wordpress.org/?v=2.5.1 Comment on HASH GROUP BY can give wrong result in Oracle 10.2 by Dinesh http://www.oracloid.com/2006/05/hash-group-by-can-give-wrong-result-in-oracle-102/#comment-62622 Dinesh Fri, 30 Jul 2010 17:23:48 +0000 http://blog.oracloid.com/?p=6#comment-62622 <a href="#comment-17784" rel="nofollow">@Alex Fatkulin </a> I remember since Oracle 5 a group by always displays the data in sorted way...if i am not wrong...but not sure if Oracle has documented that group by would return the data set sorted!!any way thanks alex.. @Alex Fatkulin
I remember since Oracle 5 a group by always displays the data in sorted way…if i am not wrong…but not sure if Oracle has documented that group by would return the data set sorted!!any way thanks alex..

]]> Comment on v$object_usage empty? by Roberto http://www.oracloid.com/2006/05/vobject_usage-empty/#comment-62274 Roberto Mon, 05 Jul 2010 13:34:06 +0000 http://blog.oracloid.com/?p=12#comment-62274 Mark, thanks for your information but I have not found the patch 67698910. Are you sure that is the right number? However, the "index monitoring usage" can be dangerous, look at: http://richardfoote.wordpress.com/2008/09/12/index-monitoring-and-foreign-keys-caution-kid-a/ Mark,
thanks for your information but I have not found the patch 67698910. Are you sure that is the right number?
However, the “index monitoring usage” can be dangerous, look at: http://richardfoote.wordpress.com/2008/09/12/index-monitoring-and-foreign-keys-caution-kid-a/

]]> Comment on v$object_usage empty? by mark teehan http://www.oracloid.com/2006/05/vobject_usage-empty/#comment-62268 mark teehan Mon, 05 Jul 2010 07:36:49 +0000 http://blog.oracloid.com/?p=12#comment-62268 If you apply patch 67698910 om 10.2.0.4 then dbms_stats (and analyze) operations will not modify the usage status during object monitoring. This patch is necessary to use object monitoring on 10g. It is fixed in 10.2.0.5. If you apply patch 67698910 om 10.2.0.4 then dbms_stats (and analyze) operations will not modify the usage status during object monitoring. This patch is necessary to use object monitoring on 10g. It is fixed in 10.2.0.5.

]]> Comment on HASH GROUP BY can give wrong result in Oracle 10.2 by Articles about _gby_hash_aggregation_enabled false volume 2 « Article Directory http://www.oracloid.com/2006/05/hash-group-by-can-give-wrong-result-in-oracle-102/#comment-61855 Articles about _gby_hash_aggregation_enabled false volume 2 « Article Directory Sat, 12 Jun 2010 14:20:06 +0000 http://blog.oracloid.com/?p=6#comment-61855 [...] Oracloid Blog ” HASH GROUP BY can give wrong result in Oracle … Hit bug 4604970 in our DW environment. This is really bad one and you might not notice that your query … are available and workaround is _GBY_HASH_AGGREGATION_ENABLED=FALSE. … [...] [...] Oracloid Blog ” HASH GROUP BY can give wrong result in Oracle … Hit bug 4604970 in our DW environment. This is really bad one and you might not notice that your query … are available and workaround is _GBY_HASH_AGGREGATION_ENABLED=FALSE. … [...]

]]> Comment on HASH GROUP BY can give wrong result in Oracle 10.2 by Articles about _gby_hash_aggregation_enabled volume 1 « Article Directory http://www.oracloid.com/2006/05/hash-group-by-can-give-wrong-result-in-oracle-102/#comment-61854 Articles about _gby_hash_aggregation_enabled volume 1 « Article Directory Sat, 12 Jun 2010 14:19:44 +0000 http://blog.oracloid.com/?p=6#comment-61854 [...] Oracloid Blog ” HASH GROUP BY can give wrong result in Oracle … Hit bug 4604970 in our DW environment. This is really bad one and you might not notice that your query … are available and workaround is _GBY_HASH_AGGREGATION_ENABLED=FALSE. … [...] [...] Oracloid Blog ” HASH GROUP BY can give wrong result in Oracle … Hit bug 4604970 in our DW environment. This is really bad one and you might not notice that your query … are available and workaround is _GBY_HASH_AGGREGATION_ENABLED=FALSE. … [...]

]]> Comment on v$object_usage empty? by Roberto http://www.oracloid.com/2006/05/vobject_usage-empty/#comment-61673 Roberto Tue, 01 Jun 2010 12:02:41 +0000 http://blog.oracloid.com/?p=12#comment-61673 <a href="#comment-61164" rel="nofollow">@Alex Gorbachev </a> Sorry, but is not in contradiction with the previous posts that you answered to Olga Yudin? However, I tried and gathering statistics set (rightly) V$OBJECT_USAGE.USED = YES @Alex Gorbachev

Sorry, but is not in contradiction with the previous posts that you answered to Olga Yudin? However, I tried and gathering statistics set (rightly) V$OBJECT_USAGE.USED = YES

]]> Comment on Bind Variable Peeking with no Histograms by Alex Gorbachev http://www.oracloid.com/2006/07/bind-variable-peeking-with-no-histograms/#comment-61165 Alex Gorbachev Thu, 29 Apr 2010 23:26:42 +0000 http://blog.oracloid.com/2006/07/bind-variable-peeking-with-no-histograms/#comment-61165 <a href="#comment-60891" rel="nofollow">@Mukesh Sharma </a> Not sure why you see it. Sorry, I can't verify this at this point. <a href="#comment-61036" rel="nofollow">@Rakesh </a> Well, yes - I mentioned it happens with 9i as well. Regarding 11g, well, let's say that adaptive cursor sharing is designed to work around the problem to a large degree but it has its own perks it seems. @Mukesh Sharma
Not sure why you see it. Sorry, I can’t verify this at this point.

@Rakesh
Well, yes - I mentioned it happens with 9i as well. Regarding 11g, well, let’s say that adaptive cursor sharing is designed to work around the problem to a large degree but it has its own perks it seems.

]]> Comment on v$object_usage empty? by Alex Gorbachev http://www.oracloid.com/2006/05/vobject_usage-empty/#comment-61164 Alex Gorbachev Thu, 29 Apr 2010 23:19:08 +0000 http://blog.oracloid.com/?p=12#comment-61164 <a href="#comment-61159" rel="nofollow">@Olga Yudin </a> Index usage is tracking whether an index been used by a user query. Stats gathering has nothing to do with index usage. Well, not directly. @Olga Yudin
Index usage is tracking whether an index been used by a user query. Stats gathering has nothing to do with index usage. Well, not directly.

]]> Comment on v$object_usage empty? by Olga Yudin http://www.oracloid.com/2006/05/vobject_usage-empty/#comment-61159 Olga Yudin Thu, 29 Apr 2010 16:51:58 +0000 http://blog.oracloid.com/?p=12#comment-61159 Hi all, I just tested running GATHER_STATS on a table and index stayed unused(NO didn't change to YES). Am I missing something? I did use cascade=>true, so Oracle analyzed table and indexes using DBMS_STATS. I used 10.2.0.4 database Hi all,
I just tested running GATHER_STATS on a table and index stayed unused(NO didn’t change to YES). Am I missing something? I did use cascade=>true, so Oracle analyzed table and indexes using DBMS_STATS. I used 10.2.0.4 database

]]> Comment on Bind Variable Peeking with no Histograms by Rakesh http://www.oracloid.com/2006/07/bind-variable-peeking-with-no-histograms/#comment-61036 Rakesh Tue, 20 Apr 2010 05:44:50 +0000 http://blog.oracloid.com/2006/07/bind-variable-peeking-with-no-histograms/#comment-61036 bind peeking happens with and without histogram also till 10g, from 11g onwards with adaptive cursor sharing this problem resolved. bind peeking happens with and without histogram also till 10g, from 11g onwards with adaptive cursor sharing this problem resolved.

]]>