Me going upstream against Naigara fall.
-Original Message-
Beilstein
Sent: Tuesday, March 27, 2001 3:19 PM
To: Multiple recipients of list ORACLE-L
Water flowing uphill (that never works either)
>>> [EMAIL PROTECTED] 03/27/01 02:52PM >>>
Turtles walking through peanut butter in January
Metalink, hands down.
-Original Message-From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]]On Behalf Of Mohan, RossSent:
Tuesday, March 27, 2001 2:52 PMTo: Multiple recipients of list
ORACLE-LSubject: OFFTOPIC - RE: Which is faster, Metalink
or...
Turtles walking through
Water flowing uphill (that never works either)
>>> [EMAIL PROTECTED] 03/27/01 02:52PM >>>
Turtles walking through peanut butter in January?
-Original Message-
Sent: Tuesday, March 27, 2001 2:32 PM
To: Multiple recipients of list ORACLE-L
A snail going uphill through jungle vegetation?
Especially when it's a stampeding herd of turtles!!!
-Original Message-
Sent: Tuesday, March 27, 2001 1:52 PM
To: ORACLE-L
Cc: MohanR
Turtles walking through peanut butter in January?
-Original Message-
Sent: Tuesday, March 27, 2001 2:32 PM
To: Multiple recipients of list ORA
Turtles walking through peanut butter in January?
-Original Message-From: Chuck Hamilton
[mailto:[EMAIL PROTECTED]]Sent: Tuesday, March 27, 2001 2:32
PMTo: Multiple recipients of list ORACLE-LSubject: Which
is faster, Metalink or...
A snail going uphill through jungle
A snail going uphill through jungle vegetation? Molasses in the northernmost region of the Siberian tundra?Do You Yahoo!?
Yahoo! Mail Personal Address -
Get email at your own domain with Yahoo! Mail.
Thanks guys.
-Original Message-
Sent: 26 March 2001 18:26
To: Multiple recipients of list ORACLE-L
Direct patch sqlldr (and insert-append) do allow
indexes - its just that you get slugged a little more
on rollback and redo...
I don't have any metrics to compare - typical usage of
eithe
Direct patch sqlldr (and insert-append) do allow
indexes - its just that you get slugged a little more
on rollback and redo...
I don't have any metrics to compare - typical usage of
either tends be to the ol' a) drop ind, b) load, c)
redindex
hth
connor
--- Martin Kendall <[EMAIL PROTECTED]> w
As far as I remember direct load allows to load data into indexed
table - it "disables" indexes for the loading time and builds the
indexes once, after the load completes. Also it prepares binary
array corresponding to the table layout and writes directly to the
file. Conventional load basical
I know that Direct Path of sqlldr does not allow Indexes
so what is the comparative performance of this suggestion if
the given Table is indexed ?
Martin
-Original Message-
Sent: 23 March 2001 09:05
To: Multiple recipients of list ORACLE-L
If you're on 8.0 or higher, try
insert /*+ AP
Hi Yong,
I don't think that's right. Try it on a table with no indexes, and dump
the redo and undo blocks afterwards. My tests show that there is no row
level redo (layer 11) except against the data dictionary tables for
space management, regardless of whether the table or tablespace is
defined a
Hi, Connor,
The append hint to insert does not disable generating rollback info. It does
stop redo generation for a nologging table.
Yong Huang
[EMAIL PROTECTED]
you wrote:
If you're on 8.0 or higher, try
insert /*+ APPEND */
into table
select * from other_table;
where "table" is defined as
Hi, Mandar,
I think your ksh is not Version 93. Here's from UWin (Korn shell for Windows,
www.research.att.com/sw/tools/uwin/):
$ var=tester
$ echo $var
tester
$ echo ${var//e/o}
tostor
Unfortunately, it's not easy to get ksh93 on a UNIX box due to (I believe)
licensing issues.
Yong Huang
[EMA
Subject:Re: Which is faster??
Thank you all for the reply. Probbaly I need to do
more test. My concern is that whether it's ok to do
the buld insert of 9 million records(say 2.7GB) on a 1
GB RBS? I think the RBS should also be at least 3GB,
Thank you all for the reply. Probbaly I need to do
more test. My concern is that whether it's ok to do
the buld insert of 9 million records(say 2.7GB) on a 1
GB RBS? I think the RBS should also be at least 3GB,
right?
Thanks,
Chris
--- Connor McDonald <[EMAIL PROTECTED]> wrote:
> If you're on 8
Chris,
Does the destination table have to be recoverable? If not you could built the table
with the UNRECOVERABLE clause and your problem is solved.
In large inserts from one table to another for archive purposes, I get around the
large RBS problem by using a procedure with a commit every 2000
If you're on 8.0 or higher, try
insert /*+ APPEND */
into table
select * from other_table;
where "table" is defined as nologging. Then you won't
hit either redo logs or rollback segments..Its the
equivalent of a sqlldr direct load
hth
connor
--- CC Harvest <[EMAIL PROTECTED]> wrote: > I have
You can try the Bulk insert feature of Oracle. It is available from Oracle 8i
onwards. It really reduces time.
--
On Thu, 22 Mar 2001 18:35:21
CC Harvest wrote:
>I have the following scripts:
>
>insert into table
>select * from table2
>;
>
>So if use the about bulk statement in my
>applicat
I have the following scripts:
insert into table
select * from table2
;
So if use the about bulk statement in my
application, and the table2 is big, say 10
million records, my concern is that it's
going to fail because of the possible rollback
segments failure. So then I have to use PL/SQL
to c
I don't think there is any question.
A smart select statement always tends to be a better solution.
The one instance where I'd definitly prefer a procedure over a select is
when the select contains Oracle's tree-walk method (i.e. connect by prior...
start with)
Regards
JL
-Original Message
Hi Gurus !
I'm going to some tables with huge amount of records. There are references between
these tables. The
question is:
Does it worth creating a procedure with several small selects or is it faster to use
one select with
joins?
For example:
CUSTOMER(CUST_ID);
CONTRACT(CONTR_ID,CUST_ID);
C
21 matches
Mail list logo
