Take the 2-minute tour ×
Database Administrators Stack Exchange is a question and answer site for database professionals who wish to improve their database skills and learn from others in the community. It's 100% free, no registration required.

Recently, i have been handed over with an ash report from DBA.

To me this report is like french. I have no idea what it is about and what all is written in this report. Can someone please guide me in reading it and explain me what all steps should i take to make my query stable and less cpu consumer. Moreover, which query is taking more of CPU.
1. Please recommend what actions should i perform.
2. Which query is consuming more of my CPU and what should i do to improve it.

share|improve this question
    
Do not ask the same question (stackoverflow.com/questions/9571246/oracle-ash-report) on multiple Stack Exchange sites. Pick one, and let the moderators determine where it belongs. –  Adam Musch Mar 7 '12 at 17:23

1 Answer 1

up vote 4 down vote accepted

I'll begin by saying that an Oracle DBA should know what the problem was by looking at ASH and AWR reports. Without AWR reports it's hard to know what the real problem is here. However, I'll help you with the first query.

The top query in the Top SQL with Top Row Sources shows the the query with the sql_id of 8t441yd5bwygd is performing a table scan. You can see the text of the query in the Complete List Of SQL Text section.

The main problem here is that there is a LIKE clause in the query, which also uses the lower() function to convert the column data into lowercase, which is then compared with the bind variable :1, which is the value passed into the query from the application.

Cut down:

select * from PERSON_CARD where lower(PERSON_ID) like :1;

To answer the question properly we really need more information regarding the data types of the columns involved & the possible values that can be passed into the WHERE clause by the application, but there are a few possibilities.

My worry is that this SQL (and probably the schema design) is generated by Hibernate, and that PERSON_ID is a long hex string UDID rather than an integer.

1) The data in the PERSON_ID column is a non-number string:

If (and only if) the input value :1 only contains a single wildcard at the end of the string (eg: LIKE "foo%") you can create a functional index on the PERSON_ID column:

create index person_card_pid_idx on person_card(lower(person_id));

This will cause an index range scan rather than a table scan (I hope!).

2) The data in the PERSON_ID column is actually an integer that exactly matches the value passed in as :1

2 possibilities here. If PERSON_ID isn't already indexed in the database, add an index. If it's already indexed, get the developers to change the application to use PERSON_ID=:1 instead of LIKE with a LOWER() call.

3) The data in the PERSON_ID column is a non-number string & the value in :1 being matched contains wildcards '%' at the beginning of the string

You'll have to use Oracle text indexes.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.