Improve company productivity with a Business Account.Sign Up

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 1356
  • Last Modified:

How to improve the select query performance for to find the count in a table in oracle

i have a huge table which consists nearly 2 million recors in oracle.
when i fire a query for to find the number for record .. it is taking very long time.

Query is : select count(*) from schema.tablename;

what should i do to improve the Performance.


Database : oracle 9.X , windows Machine.

0
ayadav1186
Asked:
ayadav1186
  • 4
  • 4
1 Solution
 
JacekMychaCommented:
Buid a bitmap index on low cardinality colum (few distinct values). Bitmap indexes have entries for null values as well, so you can pick any column. Few distinct values will keep the index small. For example, a 'GENDER' column with 3 values: M, F and NULL would be perfect.
You can also build a b-tree index on a NOT NULL  colum.
Analyze table, analyze index... You can also build index using CREATE INDEX statement with COMPUTE STATISTICS clause.

You can lookup data dictionary (SELECT TABLE_NAME, NUM_ROWS FROM ALL_TABLES ...)  to see number of rows when the table was last time analyzed. This is enough accurate for space estimates, but not good for application code.

JacekMycha
0
 
actonwangCommented:
>>select count(*) from schema.tablename;

(1) do you have index on this table? if you have index, the query will likely to access only index to get better performance instead of full table scan on table data.

(2) This query usually will use FTS ( full table scan). FTS will scan till HWM. In order to let FTS scan less blocks, you could do is to
      1.  reorgnize your table to lower blocks used thus HWM.
      2. You could use "parallel query" to improve the query performance.


acton      
     
       
0
 
Mark GeerlingsDatabase AdministratorCommented:
Why do you want to do a "select count(*) from schema.tablename" on a large table?  Do you need an exact count that is up-to-date?  Or, could you analyze the table at regular intervals (like at night or on the weekend) and just do:
select num_rows from all_tables where owner = '[schema]' and table_name = '[table_name]'

or:
select table_name, num_rows from all_tables
where owner = '[schema]'
0
Keep up with what's happening at Experts Exchange!

Sign up to receive Decoded, a new monthly digest with product updates, feature release info, continuing education opportunities, and more.

 
actonwangCommented:
to markgeer:

num_rows  is popluated only after you do statistics analysis. Also,  It might not be up to date.

Therefore I don't think you can use it to substitue for  "select count(*) ..."
0
 
Mark GeerlingsDatabase AdministratorCommented:
To actonwang:

Did you notice that I only mentioned using num_rows after saying: "...analyze the table at regular intervals...".
0
 
actonwangCommented:
to markegeer,

     >>it might not be up to date.

     did you see that?
0
 
Mark GeerlingsDatabase AdministratorCommented:
Right, the value for "num_rows" will only be as up-to-date as the last time that the table was analyzed.

(Also, I had asked the questioner if he/she needs an up-to-date count and we are still waiting for an answer to that.)
0
 
actonwangCommented:
heahea :)
0
 
Mark GeerlingsDatabase AdministratorCommented:
If you have a reporting database (a data warehouse) then a bit-mapped index may be a good option.  *DO NOT* create a bit-mapped index though on a table used in a transaction-processing system by multiple users!
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

Join & Write a Comment

Featured Post

Keep up with what's happening at Experts Exchange!

Sign up to receive Decoded, a new monthly digest with product updates, feature release info, continuing education opportunities, and more.

  • 4
  • 4
Tackle projects and never again get stuck behind a technical roadblock.
Join Now