I have the following query, which is very slow, and I tried using hint but I've seen no improvement. 
 Select 
 (sum(nvl(nu_year_open_balance,0)*nvl(b.nu_eval_rate,0)))
   from
 mst_item b,
  mst_month_summary a
  where
   b.vc_comp_code = :comp_code  and
   a.vc_comp_code=b.vc_comp_code and
   a.vc_item_code=b.vc_item_code and
   nvl(a.nu_year_open_balance,0)<>0 and
    b.vc_item_group=:item_group and
    dt_fin_start_date=:start_date

Mst_item has 4 lakhs rows and mst_month_summary has 8 lakh rows.

There is primary key index on mst_item(vc_comp_code,vc_item_code) and on mst_month_summary it's on (vc_comp_code,vc_item_code,dt_fin_start_date,dt_fin_end_date)

Is there any option to optimize the performance of this query. I tried using materialized view but it gives some error.

    Requires Free Membership to View

    Login

In the where condition above, change
b.vc_comp_code = :comp_code
to
a.vc_comp_code = :comp_code
[as a --> references mst_month_summary table and it has 8,000,000 rows. (1 Lakh --> 100,000 if I am not mistaken :)].

Similarly change
b.vc_item_code = :item_group
to
a.vc_item_group = :item_group.

So, your new where condition should be as follows: 

   a.vc_comp_code = :comp_code  and
   a.vc_item_group=:item_group and
   dt_fin_start_date=:start_date
   b.vc_comp_code=a.vc_comp_code and
   b.vc_item_code=a.vc_item_code and
   nvl(a.nu_year_open_balance,0)<>0

Let me know how it comes out. Also, please let me know which version of the database you are running.

P.S. Ensure that the tables are analyzed if you are using a cost-based optimizer. It is recommended that you use DBMS_STATS.GATHER_SCHEMA_STATS procedure to analyze the schema as opposed to only these two tables.

This was first published in March 2004

Join the conversationComment

Share
Comments

    Results

    Contribute to the conversation

    All fields are required. Comments will appear at the bottom of the article.
    Back to top