- 05-30-2012 #1Junior Member
- Join Date
- 08 Feb 2011
The first reading of data is carried out very slowlyDear
It might be better question how to keep the execution plan for the upgrade Oracle from 10g to 11g
I made uprade Oracle from 1og to 11g on HP-UX system.
ALL the first readeing of data carried out very slowly
The selected table have 10,892,769 records.
I'm executing SQL SELECT COUNT (*) FROM MyTABLE;
The first implementation of this SQL statement carried 8 min and 12 sec
>> COUNT (*) 10,892,769 1 row selected. 8:12
The another implementation of the same SQL statement carried 5 sec.
>> COUNT (*) 10,892,769 1 row selected. 00:05
What can I do to make the first implementation of the SQL statement carried only 5 sec.
Last edited by firstname.lastname@example.org; 05-30-2012 at 08:55 AM.
- 05-30-2012 #2
Re: The first reading of data is carried out very slowlyHave you tried to implement same query as select count(*) from MyTABLE; or any other query at another implementation.Regards,
- 05-30-2012 #3Junior Member
- Join Date
- 08 Feb 2011
Re: The first reading of data is carried out very slowlyI made upgrade Oracle from 10g to 11g.
After upgrade I don' have execution plan!
All first access to any tables carried out very slowly, all other accesses are o.k.
- 05-30-2012 #4
Re: The first reading of data is carried out very slowlyDear Hine,
Check out whether those tables having proper indexes or not and also whether they enabled or not.Regards,
- 05-30-2012 #5
Re: The first reading of data is carried out very slowlyAn index won't cause the second query to run in no tiime at all when the first query took such a long time. I've answered this question before in another thread and explained that 11g implements a results cache which definitely speeds up repeated access to a table given that no changes were made between the times of the two or more queries. The cache must be populated before the benefits are seen; for a first run of a query it will take as long as it takes. Yes, an index can help speed access to the data but a count(*) query will do a full table scan in the absence of a primary key index and if there are no unique keys in the data a primary key cannot be created outside of adding a column to the table, populating it with unique values and using it for the key.David Fitzjarrell
- By beaujest in forum SQL PL/SQLReplies: 5Last Post: 07-12-2011, 08:06 PM
- By Mr.777 in forum Oracle Forms and ReportsReplies: 2Last Post: 03-15-2011, 04:39 AM
- By Mamatha in forum Oracle Webcenter Suite (formerly Oracle ECM)Replies: 0Last Post: 08-25-2010, 07:22 AM
- By Nidhi25 in forum SQL PL/SQLReplies: 0Last Post: 04-07-2010, 07:12 AM
- By debrah.h48 in forum External Resources / Market PlaceReplies: 0Last Post: 12-04-2009, 08:16 AM