Question

This is a new issue that I haven’t run into before.

I have a table that at one point contained over 100k records, it’s an event log for a dev environment.

it took up to 10 seconds to load the table (simply clicking on it to view the data in the table).

I removed all but 30 rows and it still takes 7 seconds to load.

I’m using Toad, and it gives me a dialog box that says “Statement Processing…”

Any ideas?

The following are some select statements and how long they took

select * from log;  21 rows in 10 secs
select * from log where id = 120000; 1 row in 1 msec
select * from log where user = 35000; 9 rows in 7 sec

The id is the pk, there is no index on the user field.

I have a table view that contains all of the fields sitting ontop of this table as well and it runs just as slow.

Answer 1

If you issue a “select * from event_log_table”, then you are scanning the entire table with a full table scan. It has to scan through all allocated segments to see if there are rows in there. If your table once contained over 100K rows, then it has allocated at least the amount of space to be able to hold those 100K+ rows. Please see: http://download.oracle.com/docs/cd/B19306_01/server.102/b14231/schema.htm#sthref2100

Now if you delete rows, the space is still allocated to this table, and Oracle still has to scan all space. It works like a high water mark.

To reduce the high water mark, you can issue a TRUNCATE TABLE command, which resets the high water mark. But then you’ll lose ALL rows.

And there is an option to shrink the space in the table. You can read about it and its preconditions here:
http://download.oracle.com/docs/cd/B19306_01/server.102/b14200/statements_3001.htm#sthref5117

Regards,
Rob.