[Date Prev][Date Next][Thread Prev][Thread Next]
[Author Index]
[Date Index]
[Thread Index]
[SQR-USERS Info]
[SQRUG Home Page]
RE: [sqr-users] Insert -SQR
- Subject: RE: [sqr-users] Insert -SQR
- From: "George Jansen" <GJANSEN@aflcio.org>
- Date: Tue, 05 Nov 2002 09:11:35 -0500
- List-id: This list is for discussion about the SQR database reporting language from Brio Software. <sqr-users.sqrug.org>
1. A commit will close FOR UPDATE cursors. It does not affect other cursors.
2. You have not indicated--or at least I have missed--how large your insert is
compared to the size of your table. If it is a large enough fraction, then,
yes, you could save a lot of time by writing to a temporary table, disabling
indexes, and putting
ALTER SESSION SKIP DISABLED INDEXES;
into a begin-sql paragraph, and adding an /*+ APPEND */ hint to your insert
statement.
3. You really need to know where the time is going. Why not have your DBA help
you set up tests as follows. Start with a table of records such as will be
ultimately inserted into the target table.
a. Time the selection of rows from your source table. Does that finish in
reasonable time?
b. Time a single-statement insert of all rows. Does that finish in reasonable
time? If it does,
c. Time a row-by-row insertion, i.e
FOR tr IN (SELECT * FROM test_table) LOOP
INSERT INTO target_table (a, b, c, ...)
VALUES (tr.a, tr.b, tr.c, ...);
END LOOP;
/
Does that finish in reasonable time?
Oracle has tools to help you determine where the time goes. Look at
www.miracleas.dk, look at www.hotsos.com for some pointers on using the wait
interface to tune.
_______________________________________________
sqr-users mailing list
sqr-users@sqrug.org
http://www.sqrug.org/mailman/listinfo/sqr-users