__
Do You Yahoo!?
Yahoo! Health - your guide to health and wellness
http://health.yahoo.com
--
Please see the official ORACLE-L FAQ: http://www.orafaq.com
--
Author: Daniel Fink
INET: [EMAIL PROTECTED]
Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051
ORACLE-L FAQ: http://www.orafaq.com
--
Author: Daniel Fink
INET: [EMAIL PROTECTED]
Fat City Network Services-- (858) 538-5051 FAX: (858) 538-5051
San Diego, California-- Public Internet access / Mailing Lists
Greg,
The pattern with 9i was Solaris and HP-UX first, Windows next (seemed to be delayed
several months), then Linux.
Dan Fink
Loughmiller, Greg wrote:
does anyone have a reasonable idea about which Hardware platforms and their order
that Oracle releases the RDBMS??? For example,
is read-only guarantees that all data
contained within has been committed. This solution was suggested to me by a very wise
member of
the list, so it is worth a try.
Daniel Fink
BTW, AUM is not a requirement for Flashback Query. However, it is the only
configuration Oracle will support for FBQ
Rachel,
It is not active transactions against that tablespace, it is active
transactions. Yup, period! As soon as all the active transactions complete, the
tablespace will complete altering itself.
Dan
Rachel Carmichael wrote:
there WERE no active transactions against that
, there is an after-update trigger that updates tableB in the RO tablespace.
Will Oracle know this at the PARSE phase?
Daniel Fink
Mark Richard wrote:
I think Daniel covered this when he mentioned that a transaction can
consist of many statements - some of which Oracle may not yet be aware of.
Until a user
to write sort runs to disk (temporary tablespace). Is this what they mean by
swapping?
The best approach (IMHO) is to go back to the critics and ask for more
detailed information, including the logic and scenarios describing the logic.
Daniel Fink
David Turner wrote:
I was presenting
think you will have to recreate the database
(since it is the system tablespace). If it were another tablespace, I think you could
get by with just recreating the tablespace.
Daniel Fink
-Original Message-
From: Gurelei [mailto:[EMAIL PROTECTED]
Sent: Thursday, June 19, 2003 1:55
I figured he try for something closer to home...say Oragon?
Then Bill would want to buy Mississippi (postal abreviation MS for our non-us listers).
Jesse, Rich wrote:
Oracle Makes Bid to Buy State of Delaware
Software giant willing to assume state's $225 million deficit in exchange
for
Another thing to consider is that the object id may be needed for
recovery/rollback/read consistency purposes. The object id is one of the fields kept
in the undo record. I presume this is for a sanity check to make certain that the undo
is being applied to the right object. I wonder if the
It depends on the access path and any other implicit sorting.
Access path - on a full table scan, the data is returned in the physical order it is
stored in the object. An index scan will return the data sorted according to the rule
of the index.
Implicit sorting - distinct will cause an
I am reviewing a 10046 trace file and have come across several questions (actually
more like hypothesis that I want to confirm).
1) The first user statement executed waits only on SQL*Net to/from client and does not
wait on any db file activity. I presume this means that the data was found in
David,
Have you tried removing the around the TO_DATE call? The indicate a
character string and Oracle is unable to translate that string into something
acceptable for a DATE datatype.
David Lewandowski wrote:
Thanks Lisa. But regrettably my LOADDATE isn't always SYSDATE. Do you
Run the following command after both tables are loaded and the constraint is not
enabled
select vs_off_sysid
from cms.twesvs
where vs_off_sysid not in (select of_off_sysid from cms.twesof);
[EMAIL PROTECTED] wrote:
Yes, I am sure. Just to make sure, I have compared the counts from export
When all else fails...
Log a TAR with Oracle
Sit back and wait, while enjoying a few belts of quality Kentucky sipping Bourbon
Read through the phone book and locate all the Ursulas in the greater Frankfort area...begin:vcard
n:Fink;Daniel
tel;cell:303.808.3282
tel;work:303.272.3225
Each row that is retrieved from the database that matches the WHERE condition is
placed in a result set. The sequence within the result set is the value in the
pseudocolumn ROWNUM. If Oracle is using a full-table scan, the first row of the first
block is the first in the result set and is
IF you are looking to break up the update into more 'manageable' pieces, here is a
thought...
UPDATE EMP2
SET GENDER = 'F'
WHERE GENDER = ' '
AND ROWNUM = 100;
UPDATE EMP2
SET GENDER = 'F'
WHERE GENDER = ' '
AND ROWNUM = 100;
*Since the first 100 records have already been updated, the second
Repeat after me...readers never block writers...writers never block readers...
The update will lock the row(s) of the result set, but it will not block anyone from
reading the data. However, they may not read the modifications made by the update
until it is committed.
MaryAnn Atkinson wrote:
Depending upon the version, you may be able to use the analytical functions (8.1.6+)
to achieve what you want. I don't have a ready example, but the Oracle docs are pretty
good. If you want a less elegant solution, use an inline query, sort by the reverse
order (asc or desc depends on your
I'll agree with Rachel's methodology and add another consideration.
Look at separating constraint indexes (primary keys, unique, perhaps even foreign
keys) from performance indexes. If you find resource constraints on backups
(time/disk), you can safely ignore the performance indexes. The
-aid solution, but it allowed us to focus on backing up
the critical components successfully.
Daniel Fink
Mercadante, Thomas F wrote:
I disagree with the concept of recovery not including some indexes because
they can be rebuilt later. To me, that's like going to a gas station and
only
I may be way off base here, so any gurus please correct me with a gentle slap to the
back of the head...
Index and table access is not as simple as index entry..table row..index entry..table
row..etc. I just ran a quick test (which may not be represntative and is using the
primary key which
In one case, the statement's execution plan used a nested loop, where it read the
local table then probed the index on the remote table. This caused several million
trips across the network to retrieve less than 10,000 records (IIRC).
select local.cola, local.colb, remote.colc, remote.cold
from
There's a good argument for separating tables/indexes.
[EMAIL PROTECTED] wrote:
If you ran dbverify with the database up, it's very likely that
there is nothing wrong with the datafiles.
If dbv reads a block in transition, it will appear corrupt.
Run it again, and there will either be
Kirtikumar Deshpande wrote:
OracleWorld web sites SFO/PARIS revealing the new name as: Oracle10G
Gee, it's not 'i' anymore ?
Nope!! It's Capital G.
G for God?
No. It's Grid :)
Enjoy..
- Kirti
The first thing that comes to mind (as a commuter) is Grid...lock! I can see it
now...New
I just realized what the upgraded name for 9ias will be...begin:vcard
n:Fink;Daniel
tel;cell:303.808.3282
tel;work:303.272.3225
x-mozilla-html:TRUE
adr:;;
version:2.1
email;internet:[EMAIL PROTECTED]
title:DB Services Lead
x-mozilla-cpt:;-4832
fn:Daniel Fink
end:vcard
How about inserting 'NULL'? No, I am not kidding... The following is taken from an
actual system...
columnA has a NOT NULL constraint
columnA has a CHECK constraint with a list of acceptable values, one of them is 'NULL'
BTW, I don't recommend either approach. NULL has a special meaning to
I don't think SHOW is a function that you will find in any of the Oracle db creation
scripts. It is most likely part of the sql*plus program. The keyword SHOW can be used
for a lot of things (USER and PARAMETER come to mind). Obviously, it reads some data
structures, but I am sure it is not a
Michael,
When you say 'different execution plans', what are the differences? Is the
choice in execution plans causing the ORA-07445? If so, why are the execution plans
different (there has to be a reason)? Statistics, parameters, etc. will alter the
execution plan choices.
The
Okay, so its Monday...
What I meant to say is that are the queries with different execution plans the ones
encountering 7445s. If there is a common thread in the execution plans (i.e. hash join
instead of nested loops), then the hj may be triggering the 7445s.
Actually, 10g will have a new
SHOW PARAMETER performs a 'FIXED TABLE FULL' scan of x$ksppi and x$ksppcv. Exactly
what I would expect.
SHOW USER did not show up in the trace file, which tells me it is internal code path.
SHOW ALL did the same. I think that the difference is that PARAMETERs are database
level information,
Ron,
Are the sizes of the data manipulated by the batch load the same?
Is the db_block_size the same on each database?
What is the size of the rbs during processing? What is the hwmsize, shrinks,
extends from v$rollstat?
You may be encountering a situation where
Jared,
If you will closely examine the specs, the assumption is not documented.
Therefore, the application code must take into account that the 1st Wednesday of the
month may still be in the future.
select CASE WHEN to_char(to_date('08/07/2003','MM/DD/'), 'DD') = 6
THEN
I was perusing a 10046 trace file and I noticed that timestamps are written to the
trace file. Sometimes they were very regular (3 minutes apart give or take 30 seconds)
while other times they were hours apart. I have noticed that two timestamps are never
written without any intervening
Metalink has a note that the INLIST ITERATOR may cause problems on certain platforms.
That's about as much detail as it offers.
I don't know the internal workings of II, but I suspect there is some sort of merging
occuring after each (or all) iterations. Perhaps someone on the list can explain
Phillipe,
Look at using INSTR and SUBSTR to calculate the position of the last / and
work from there.
Daniel
NGUYEN Philippe (Cetelem) wrote:
Hi Gurus!
a very simple problem for You :I just want to retrieve the .dbf name from file_name
column in dba_data_files.
eg
-Original Message-
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Daniel Fink
Sent: 22. júlí 2003 21:19
To: Multiple recipients of list ORACLE-L
Subject: Timestamps in trace files
I was perusing a 10046 trace file and I noticed that
timestamps are written
Building on Jacques' solution, here is one that aligns all the statements nice and
neat. Of course, it does not have the username/instance, so I don't know how
acceptable it is
select ' 1 '
as new_prompt
from dual
/
set sqlprompt new_sqlprompt
1 select 1,
2 2,
3 3,
4 4,
David,
This comes with the usual caveats (contact oracle support, don't do this
without testing, some fading may occur, objects in the rear view mirror are larger
than they appear)...
On error dumps are pretty straight forward. If you can reproduce the error in
a session, just
Oracle will not let you drop a rollback segment if there are active transactions using
it. However, it will allow you to offline the segment and no new transactions can use
it. I don't recall the exact status in v$rollstat, but I think it may say pending
offline.
As for the commit across
As much as I hate to disagree with the DBA Goddess...
The database will not HANG. A more appropriate term is that further processing is
halted until it is able to archive the redo log before it begins to overwrite it.
You will receive a message in the alert log, I don't recall the exact text,
Henry,
I'll make an attempt, but I am still learning a great deal about wait events
and trace files. Cary, Mogens, Anjo, Tim, Jonathan, Wolfgang, et.al. are better
authorities, so any corrections are very welcome.
The time between waits is the elapsed time. If we equate elapsed
is already reflected in the trace numbers.)
There is something very strange hear but I don't know what. Might be an
interesting case to track db/os relationships (running on AIX 4.3)
Henry
-Original Message-
Daniel Fink
Sent: Thursday, July 24, 2003 11:49 AM
To: Multiple
10046 Trace files are a peak into how an application and Oracle operate. In looking at
one generated by a web server, I identified that the application (user) was sorting to
a temporary tablespace that was dictionary/permanent without at the data dictionary. I
saw a query performing DML against
for the rbs? In v$rollstat, how many extends and shrinks are you
seeing? In terms of cause #2, what is your db_block_size, how many other
transactions are running at the same time and how many rbs do you have?
Daniel Fink
Smith, Ron L. wrote:
That's just sidestepping the problem. I agree
, but the database was not opened. I know there
are registry settings to control this, but I don't know them off the top of my head. I
am sure someone on the list can provide them.
The bottom line...imp did not close the database, the server reboot did.
Daniel Fink
Wolfe Stephen S GS-11 6 MDSS/SGSI wrote
Is the INLIST ITERATOR unable to use the index unless we specify NOT NULL? I wanted to
bounce this off the list before we log a TAR.
We are examining the performance of a query and I am trying to understand why an
INLIST ITERATOR is not used if there is not an explicit IS NOT NULL predicate
The scenario you describe is not consistent with manual undo. Was the database
created/started with automatic undo and you have switched to manual?
[EMAIL PROTECTED] wrote:
I use manual managed undo tablespace but there are some system named rollback
segments on that
tablespace ( I did
Michael Kline wrote:
There seems to be some big confusion if Veritas is
working or not...
snip...
Recovery has NEVER been tested.
Michael Alan Kline, Sr.
Principal Consultant
Business to Business Solutions, LLC
Phone: 804-744-1545 Cell: 804-314-6262
ICQ: 1009605, 975313
Email:
If you are now running in manual undo mode and have created rollback segments, you can
remove the undo segments. Verify that the segments are offline and you are not using
them. Then drop the undo tablespace. This is the only way to get rid of them.
[EMAIL PROTECTED] wrote:
yes . What should
Thanks to Wolfgang for spotting the problem. It was not the inlist iterator at all but
an outer join! The NOT NULL predicate invalidated the outer join, so the optimizer was
smart enough to make a different decision. I am still perplexed as to why the table
access information was so radically
Tanel Poder wrote:
Hi!
Statistic#'s vary between versions (and possibly platforms), so one should
use statistic names instead of numbers in scripts.
Statistic# 99 is physical reads direct (lob) in my test environment for
example.
Tanel.
An alternative is to use v$sess_io and examine
WOW! How many googolbytes do you have devoted to your Undo Tablespaces?
Jamadagni, Rajendra wrote:
TICK : Fri Aug 8 09:06:03 2003
SEARCH in kdisti: tsn = 5, objd = 83525, rdba = 33588489
ORA-01555 caused by SQL statement below (Query Duration=1060347963 sec, SCN:
0x0011.05e003c2):
TICK
Abraham,
Setting the retention time may not solve the problem. One of the ways that an ORA-1555
can be triggered is when the transaction table slot is overwritten. This is caused by
having many small, serial transactions in the database while the export is running.
In each undo segment (or
Beginning in 9, the sql kernel was incorporated into pl/sql. This means that any sql
call is now available to pl/sql. In 8.x and earlier versions, they were separate and,
therefore, some calls in sql were not available in pl/sql.
Danielbegin:vcard
n:Fink;Daniel
x-mozilla-html:FALSE
org:Sun
One of our databases has the 10235 (level 2) event set. Unfortunately, the Metalink
doc is very sparse (barely more than oerr). It mentions that it protects internal
memory management structures. What I am needing are reasons why this event would be
set. Anyone have experience with this event
Tim Gorman wrote:
I don't mean to be argumentative, but every time I see assertions like
these, I suspect someone has been reading some rather discredited books...
Okay, now I have an image of Tim, in a kilt, standing over a grate, with his face
painted with Blue vertical stripes and a big
Sherrie,
Also check out Tim Gorman's paper/presentation Cat's, Dog's and ORA-1555s on his
website www.evdbt.com. It is a great place to start, then move on to my papers which
are more geeky in nature (reading block dumps, following the undo chain, etc.).
Daniel Fink
(Gotta run, just checking
Stephen,
Tim's statement is correct, but can be construed incorrectly if you read it and think
of TEMP segments. AUM still uses undo segments (same basic structure as rollback
segments). However, one of the space management steps is to allow an undo segment to
'steal' extents from another undo
they often have some top notch meetings that are easy to attend and
afford.
Daniel Fink
[EMAIL PROTECTED] wrote:
Jared et al,
I was just told by my new boss that I have to have any requests for any
conferences/shows in by Sept. 5th (his budget deadline). It's been years since I
worked
From a technical and business perspective, what are the reasons to migrate from
32-bit to 64-bit Oracle? Are there known bugs/problems with one version that are not
present in the other?
Daniel Finkbegin:vcard
n:Fink;Daniel
x-mozilla-html:FALSE
org:Sun Microsystems, Inc.
adr:;;
version:2.1
to identify what is slow, when it is slow. If they can repeat it, run
a test with 10046 enabled. That WILL identify
the cause of the response time degradation.
Daniel Fink
[EMAIL PROTECTED] wrote:
We have two teams. Database team and a Software engineering team. The software
engineering team wrote
Cary is spot on. The technical problem is fairly easy to address. The political
problem is the real issue. Will your manager support
you in laying down the 'rules of engagement'? Do you have the ability/support to push
back and say I will be glad to help, but I
need the following from the
In testing, I found that MAXTRANS seems to be ignored in 9.2.0.3.
I created a table with a MAXTRANS setting of 1. I inserted some rows and committed. In
one session, I updated a row, but did not
commit. In another session, I updated a different row (in the same block). I expected
to see a wait,
BINGO! I tried 3 tx and #3 waited on a tx to end.
I just RTFMed and the doc (both r1 and r2) says it can be 1 to 255. If you can recall
where you read this, please let me know.
Daniel
Tanel Poder wrote:
Hi!
I remember from somewhere that in 9i, mintrans is 2 for tables as well, like
for
I was curious about your statement the kernel only peeks once per
session. I wondered if another session that executed the same statement
would be impacted by the peeking or would the optimizer reparse the
statement.
So I set up a test. Million row table with 2 columns. c1 - number, c2 -
date. C1
Guang,
What is the value of dba_tables.blocks for each of the tables?
What is the value of db_file_multiblock_read_count for each of the dbs?
Daniel Fink
Guang Mei wrote:
CLUSTERING_FACTOR values of the indexes for table IDENTIFIER are identical
on both servers. Also all table and indexes
fail
Babette Turner-Underwood wrote:
test
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: Babette Turner-Underwood
INET: [EMAIL PROTECTED]
Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San Diego, California-- Mailing list and web
The Rocky Mountain Oracle User Group needs YOU! We are looking
for a few (well, more than a few) good speakers for our annual
Training Days conference to be held in Denver, Colorado on
February 11 12, 2004. This is your chance to share your
knowledge and experience with fellow Oracle
David,
Do you have timed_statistics set to true? If not, you need to do so. Although I
am pretty certain that this is the case, but let's assume that you truly waited
0 seconds. If you were able to remove ALL of the latch free events, you would
improve the response time in the system by 0 seconds.
Let's just cut the crap and start our own Oracle-L Masters (OLM).
Qualifications
1) Willingness to answer questions
2) Willingness to ask questions, after RTFM
3) Willingness to publicly admit mistakes and offer corrections
4) Demonstration of advanced knowledge of Oracle
5) Able to back up
)
4) Instance crashes (may not record termination in alert log).
I am wondering if there is an issue with AUM and a large number
of segments. Any thoughts, experiences, etc. are greatly
appreciated.
Daniel Fink
begin:vcard
n:Fink;Daniel
x-mozilla-html:FALSE
org:Sun Microsystems, Inc.
adr
The formula is ceiling(# of digits/2) + 1 + (1 if negative).
[EMAIL PROTECTED] wrote:
How many bytes does oracle use to store number datatype ?
--
Please see the official ORACLE-L FAQ: http://www.orafaq.net
--
Author: [EMAIL PROTECTED]
INET: [EMAIL PROTECTED]
Fat City Network
You also need to consider replication issues. If you will have another
site that needs to store the documents, you will need to include that in
the decision. For example, if you store them inside, Oracle can handle
the propogation of the documents to the alternate site. If you store them
outside,
Karniotis, Stephen wrote:
Does this help everyone understand what we are looking for? SUBMIT SUBMIT
SUBMIT
shuffling and cringing
Yes, master. Whatever you wish, master.
Thank You
Stephen P. Karniotis
Technical Alliance Manager
Compuware Corporation
Direct: (313) 227-4350
Mobile:
I realize that it may be terrifying, but there is a great deal of value in becoming
indistinguishable. How many accounting/hr/finance jobs are going overseas? None that I
have heard of. There are a ton of managers I wish we could offshore (but only about 1
mile off shore...). Bear with me for a
I can see it now.
2103 - Can you believe it! They just outsourced my job to a martian!
2203 - Can you believe it! They just outsourced my job to some guy from Alpha
Centauri
2303 - Can you believe it! They just outsourced my job to a being from a parallel
dimension!
2403 - Can you believe it!
Elain,
The query is returning the correct data. DISTINCT applies to the row as a
whole. Since the NAME adds a new element, DISTINCT finds that 1MIKEY is
different than 1JOEY.
How do you determine the 'first' row of the testid? Depending upon the
access path (table/index), this could be a
Some random thoughts, with examples
1) Tracks historical information to determine when a value is out of normal
range or when a value has suddenly increased.
ex. - SQL*Net messages consume 85-95% of wait time. If this value drops to
75% or rises to 99% alarm.
- AR_DATA tablespace is
I've done a little with it. It's free, which is a plus and it parses the trace file
and loads it into
database tables. The documentation is limited to a single readme, which is adequate.
The biggest problems
with it are that it requires you to run the analyzer as the user issuing the query, it
an eye on the begin_time and end_time. However, for
estimation purposes it should work.
Daniel Fink
Thomas Jeff wrote:
I'm beginning the process of converting over to automatic
undo management. I'm wondering as to exactly how large to
initially build the UNDO tablespace.Make it roughly
you are using AUM, keep a close eye on
v$undostat, though there are some known issues with it not populating
properly, keep an eye on the begin_time and end_time. However, for
estimation purposes it should work.
Daniel Fink
Thomas Jeff wrote:
I'm beginning the process of converting over
of memory? Is there an upper bound?
Daniel Fink
begin:vcard
n:Fink;Daniel
x-mozilla-html:FALSE
org:Sun Microsystems, Inc.
adr:;;
version:2.1
title:Lead, Database Services
x-mozilla-cpt:;9168
fn:Daniel W. Fink
end:vcard
Today is the last day to submit an abstract for Rocky Mountain
Oracle User Group Training Days 2004 in sunny and snowy Denver,
Colorado.
Last year, we had sessions from oracle-lers Cary, Anjo, Rachel,
Tim as well as Stephan Haisley, Gary Goodman, Kent Graziano,
Dave Ensor, Craig Shallahamer and
the same error.
Daniel Fink
Teresita Castro wrote:
HI!!I want
to create the next index:CREATE
UNIQUE INDEX LAWSON2.IOEINVCLINE1 ON
"LAWSON2".OEINVCLINE(COMPANY,
INVC_PREFIX, INVC_NUMBER, ITEM)
TABLESPACE LAWSON_PRUEBAS_INDICES
PCTFREE 5 STORAGE(INITIAL 40960 )But
I can't because Orac
Looks like a documentation bug. The description exists in the 901 docs but not
in 920 docs. The view still exists in 920.
[EMAIL PROTECTED] wrote:
List, I am unable to find v$sort_usage in the 9i docs, though this synonym
exists. Any idea why ?
--
Please see the official ORACLE-L FAQ:
probably just do that anyway. :-)
-Original Message-
Daniel Fink
Sent: Monday, September 29, 2003 3:45 PM
To: Multiple recipients of list ORACLE-L
Today is the last day to submit an abstract for Rocky Mountain
Oracle User Group Training Days 2004 in sunny and snowy Denver,
Colorado
(1048576k /20)
indicates that your average row length is over 5k. If you have 8k blocks, this means
an average of 1 row per block (perhaps less depending on the variance in row length).
Daniel Fink
Kaing, Leng wrote:
Hello everyone,
Env: 8.1.7.4, SunOs 5.8 64 Bit
We seem to hitting bug
Henry,
What happens if you issue another query after the query of interest?
(something like "select 1 from dual;") STATshould be emitted when
the cursor is closed.
Daniel
Henry Poras wrote:
Thanks.
I tried both disabling the trace and quitting from the session. No luck
with 10046, just
Bambi,
IIRC, you need a BREAK in order for the COMP to work, even when it is
ON REPORT.
Daniel
Bellow, Bambi wrote:
Friends --
Did I miss something dying? The doc sez it's still there, but SQL*Plus
doesn't seem to be recognizing it...
col owner form a10
col segment_type form a5
col
Sinardy,
I'm not sure I would perform the export/import. It may temporarily mask the problem,
but it will return. I think the real issue is that the table has a very large average
row length (4898). If you look at the chain count (roughly 1% of the
rows), this indicates that the updates are
Johan,
First, you don't need the distinct. The proper query will return
1 row per ip. Second, take the max(timestamp) out of the group
by. That is causing the problem.
Daniel
Johan Muller wrote:
I have multiple timestamps values for single ip in a table, I
need the max(timestamp) for each
this sequence of events in
the trace file. Any assistance is greatly appreciated.
Daniel Fink
begin:vcard
n:Fink;Daniel
x-mozilla-html:FALSE
org:Sun Microsystems, Inc.
adr:;;
version:2.1
title:Lead, Database Services
x-mozilla-cpt:;9168
fn:Daniel W. Fink
end:vcard
I prefer to tell BAD data to go sit in the corner and think about why it is
BAD. After a suitable time, I'll let the BAD data out of the corner and it can
rejoin the rest of the data, as long as it agrees to play nice and not upset
the other data.
Yechiel Adar wrote:
Have you tried to run
This is not an issue of answering the question, but pointing out that the question is
not correct.
Why do databases exist (aside to make Larry money)? To 'permanently' store data. As
this storage must survive a system failure, we choose to place the data on a
non-volatile medium (disk, paper,
IIRC, you will also need to change who is the running owner of the oracle
process. In WinNT/2k, this defaults to SYSTEM. You need to create an
ORACLE user on each system and change the instance owner to ORACLE. Then deal
with the normal permissions stuff. There should be a note on metalink on how
be those half filled?
We may have to build a special 16K block database just for this table in the end.
Now I see why we need 9I :-} But upgrading is probably not an option at the moment.
Thank you!!
Leng.
---
From: Daniel Fink [EMAIL PROTECTED]
Date: Wed, 01 Oct 2003
next big project. I'll still
be hacking away at Oracle, learning how it REALLY works, doing things that they
say should/could not be done. But my goal is to make myself more proficient in
a skill that will benefit the organization.
My $0.03 (I'm a little long winded today...)
Daniel Fink
Freeman
Ah, yes... No kids here, but I do have pet potbellied pigs to support. Anyone
have more than 2?
;)
Dan
Brian Dunbar wrote:
Freeman Robert - IL [mailto:[EMAIL PROTECTED] on Friday, October 03, 2003
12:09 PM
Well I've brought my share of children into the world, thats for
sure I'd
Isee what you mean. Too funny!
Hmm...Iguess I've been wrong all these years. And all that time
Iwasted reading the official documentation on such "undocumented"
and "hidden" programs like export/import, sql*loader, tkprof...
[EMAIL PROTECTED] wrote:
OK, well I thought it was funny.
Doing a
1 - 100 of 161 matches
Mail list logo
