Link to home
Start Free TrialLog in
Avatar of tks_g100
tks_g100Flag for United States of America

asked on

SQL Query takes long time to execute -Need Help in fine tune the query.

Query takes more time to Execute. Please suggest how to improve the execution timing.


SELECT
z.end_user_emp_id_,
y.name,
y.reconciliationidentity,
y.instanceid,
y.requestid,
y.assetid,
y.classid,
e.remedy_login_id as LoginCTM,
e.person_id as personID_CTM,
e.instanceid as InstanceCTM,
e.internet_e_mail,
e.phone_number_business,
e.full_name,
e.default_notify_mechanism,
f.ENTRY_ID
From
(

select v.flag_change_end_user_emp_id_, v.asset_id_, v.modified_date, v.end_user_emp_id_ from (select flag_change_end_user_emp_id_, asset_id_, modified_date, end_user_emp_id_ from Aradmin.ams_ast_field_tracking@amsbase.itg.ti.com a
where not exists
(select 1 asset_id_ from owner.table1@EDB b
where a.asset_id_=b.asset_id_ and b.modified_date > a.modified_date )) v where v.modified_date > '1223855640' and v.flag_change_end_user_emp_id_= 0 order by v.asset_id_


) z
LEFT JOIN BASETAB y ON substr(z.asset_id_, 4) = substr(y.assetid, 4)
LEFT JOIN PEOPLETAB e ON z.end_user_emp_id_= e.remedy_login_id
LEFT JOIN ASTPEOPLETAB f ON substr(z.asset_id_, 4) = substr(f.asset_id_, 4)
ASKER CERTIFIED SOLUTION
Avatar of dro_law
dro_law
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Hello also could you provide us the table row count for each table it will help to point where it is slow

I never used oracle before, could you tell me what the table name Aradmin.ams_ast_field_tracking@amsbase.itg.ti.com  mean, is it a kind of linked server ? if yes do you have good network performance with it ?

Also do not forget that you use three nested query, if it is possible try to change those to use simple join it will improve the performance
What the above posters said - execution plan would be a great help in identifying the bottlenecks.

Also, just looking at the query, it looks like the various asset_id fields you use in your join/where criterias are of string (varchar) type, correct? Using string types as criterias are generelly very "expensive" performance-wise. Especially if they are not indexed, if the data set is large and/or if you're using functions like substr in the criteria.

Also, as was already suggested, try to modify your query to remove the nesting. Unless not possible, I generally advice to join rather than nest queries.