Question about Parallel Query
Currently running Oracle 8.1.7.3 on 32 CPU HP box. I'm doing some data analysis and ran in to some inconsistencies on what columns parallel query works on. For most columns it kicks off all the threads(ora_pxxx) but on a few of the columns, it will only run 1 thread. These queries run full table scans.
Example SQL:
Select /* parallel (column1, 10) */ column1
from table1
I can't seem to find much documentation on parallel queries.
sorry. correction on the SQL format:
Select /*+ parallel(tb1, 20)*/ column1
from table1 tb1
Similar Messages
-
Question about Parallels and ATI Radeon 5770
Ok, so I have a Mac Pro 3,1 here. I'm upgrading it to 16GB and two 7200 RPM 3TB drives which I'll run in Raid 0.
It has an ATI 2600HD with 256MB VRAM. This card has two DVI ports.
I just purchased an ATI Radeon 5770. I will use it with two Cinema Displays (DVI connections).
To connect the second display, I have the Apple Mini DisplayPort to Dual-Link DVI Adapter at Firmware 1.03.
My question concerns Parallels 8.
My Windows 7 VM will be a heavy-use VM so I want snappy performance. I was planning to dedicate 8GB RAM, half my CPU cores to it, and 512MB VRAM on the 5770.
What I was wondering about was the possibility of leaving in the 2600 HD and using its GPU to support my Parallels VM. Is that feasible? Practical?
ThanksDump the 2600XT (put it on the shelf, don't use it)
Install Windows natively first, then you can use it both ways, natively and dual boot, plus in a VM when that works.
Another option is 10.8.2 supports more Nvidia cards and more then the 1GB.
Don't use green 3TB in an array. -
Some question about the query designer
hello, dear all,
I am a new comer of here, and I am intersting in BI, but I have no basic knowledge about it.
so I just want someone could give me some advice about it.
our boss need I do the developer about the query designer, I just have searched in this forum. but nothing founded for I am a new comer here,
I heard there are some training document of the query designer, could someone give me the URL, thanks.Hi,
Query desinger is used to develop a Query, Query can be created on following data targets
-Info Cube
-DSO
Virtual data target
-MultiCube
-Infoset
-Multiprovider
We have 5 section in query designer
- Infoprovider : where we select the data target , on which report to be created
-Filter : to restrict value at infoprovider level ( if you want data for year 2008, for example)
-Free Characterstic : this allow you to drill down
-Columns : char/keyfigs to be display in columns can be added here
Row: key/char to be display in Rows can be added here
gv me ur mailid i will let u have Bex manual,
I would suggest , if you have any IDES ( training ) system , where you can log in and then go to RRMX,
and try to create new query and add any data target ( which is already created ) and then drag and drop the char/key fig to the required section ,
save it and execute it .....
if you play arround and see the output , that would help u to understand how to work with query designer.
Hope this helps
Sukhi
Edited by: Sukhvidner Singh on Nov 4, 2009 5:36 PM -
Hi all ,
I have small question related to bad query .
what is the affecting happens to database if there is lot of bad query , i know the performance issue is there what's other things .like archive log generated database time , IO , TMP tablespace , please give me information about it's .
thanks & regards.user11969912 wrote:
I have small question related to bad query . What do you consider a bad query? A bad query can be the result of poorly written and illogical SQL. It can be due to not using bind variables. It can be due to a on-optimal execution plan generated by the CBO. It can be due to poorly designed code that uses what seems to be a "good query", badly and in the wrong way (using bulk collection when a native SQL alone suffices, or hitting the very same data multiple times, etc).
what is the affecting happens to database if there is lot of bad query Each of these have a different impact on the database. A "bad query" can cause a snapshot-too-old error. A deadlock error. A shared pool memory allocation error. Can cause no error and simple increase I/O. Or increase CPU. Etc.
It is a lot more complex than what you seem to think, given your question. -
Question about the Query Optimizer
For an Excercise during my database lecture the following table
CREATE TABLE TASKS
"ID" NUMBER NOT NULL ENABLE,
"START_DATE" DATE,
"END_DATE" DATE,
"DESCRIPTION" VARCHAR2(50 BYTE)
) ;with about 1.5 Million Entries were given. In addition there was the following Query:
SELECT START_DATE, COUNT(START_DATE) FROM TASKS
GROUP BY START_DATE
ORDER BY START_DATE;And the Index:
create index blub on Tasks (start_date asc);The main excercise was to speed up queries with indexes. Because all data is accessed the optimizer ignores the index and just did a full table scan.
Here the QEP:
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
| 0 | SELECT STATEMENT | | 9343 | 74744 | 3423 (6)| 00:00:42 |
| 1 | SORT GROUP BY | | 9343 | 74744 | 3423 (6)| 00:00:42 |
| 2 | TABLE ACCESS FULL| TASKS | 1981K| 15M| 3276 (2)| 00:00:40 |
----------------------------------------------------------------------------Then we tried to force him to do the index with this query:
ALTER SESSION SET OPTIMIZER_MODE = FIRST_ROWS_1;
SELECT /* + INDEX (TASKS BLUB) */ START_DATE, COUNT(START_DATE) FROM TASKS
GROUP BY START_DATE
ORDER BY START_DATE;but again it ignored the index. The optimizer guide states clearly, that each time you access all data from a table it has to do a full scan.
So we tricked him in doing a fast index scan with this query:
create or replace function bla
return date deterministic is
ret date;
begin
select MIN(start_date) into ret from Tasks;
return ret;
end bla;
ALTER SESSION SET OPTIMIZER_MODE = FIRST_ROWS_1;
SELECT /* + INDEX (TASKS BLUB) */ START_DATE, COUNT(START_DATE) FROM TASKS
where start_date >= bla
GROUP BY START_DATE
ORDER BY START_DATE; now we got the following QEP:
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
| 0 | SELECT STATEMENT | | 1 | 8 | 3 (0)| 00:00:01 |
| 1 | SORT GROUP BY NOSORT| | 1 | 8 | 3 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | BLUB | 1 | 8 | 3 (0)| 00:00:01 |
----------------------------------------------------------------------------- So it do use the index.
Now to my two questions:
1. Why should it always do a full scan (because the optimizer documentation answer is a little bit unsatisfying)?
2. After looking at the difference between the costs (FS: 3276 IR: 3) and the time the system needs (FS: 9,6 sec IR: 4,45) why does the optimizer refuse the clearly better plan?
Thanks in advance,
Kai Gödde
Edited by: Kai Gödde on May 30, 2011 6:54 PM
Edited by: Kai Gödde on May 30, 2011 6:56 PMJohn Spencer mentioned already the most important point (the role of NULL values) but here are some minor additions:
* with the additional NOT NULL condition Oracle will do an INDEX FAST FULL SCAN (a multiblock scan of the index segment, similar to a FULL TABLE SCAN)
* with the additional NOT NULL condition you don't need a hint and Oracle will choose the IFFS access
* if the index was a bitmap index you would not need a NOT NULL condition because Oracle stores NULL-values in bitmap indexes (but you should not define bitmap indexes in an OLTP system because they bring massive locking issues)
* if the index would contain a second value the additional NOT NULL condition would also be needless
CREATE TABLE TASKS
"ID" NUMBER NOT NULL ENABLE,
"START_DATE" DATE,
"END_DATE" DATE,
"DESCRIPTION" VARCHAR2(50 BYTE)
insert into tasks
select rownum
, case when mod(rownum, 5) = 0 then null else sysdate - round(rownum/100) end
, case when mod(rownum, 4) = 0 then null else sysdate - round(rownum/50) end
, lpad('*', 50 , '*')
from dual
connect by level <= 1500000;
exec dbms_stats.gather_table_stats(user, 'TASKS')
-- the IFFS access without a hint:
create index blub on Tasks (start_date);
SELECT START_DATE
, COUNT(START_DATE)
FROM TASKS
WHERE START_DATE IS NOT NULL
GROUP BY START_DATE
ORDER BY START_DATE;
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
| 0 | SELECT STATEMENT | | 15001 | 102K| 1871 (16)| 00:00:10 |
| 1 | SORT GROUP BY | | 15001 | 102K| 1871 (16)| 00:00:10 |
|* 2 | INDEX FAST FULL SCAN| BLUB | 1200K| 8203K| 1631 (3)| 00:00:09 |
Statistiken
0 recursive calls
0 db block gets
3198 consistent gets
0 physical reads
0 redo size
378627 bytes sent via SQL*Net to client
11520 bytes received via SQL*Net from client
1002 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
15001 rows processed
-- the bitmap index
create bitmap index blub on tasks(START_DATE);
SELECT START_DATE
, COUNT(START_DATE)
FROM TASKS
GROUP BY START_DATE
ORDER BY START_DATE
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
| 0 | SELECT STATEMENT | | 15001 | 102K| 132 (1)| 00:00:01 |
| 1 | SORT GROUP BY NOSORT | | 15001 | 102K| 132 (1)| 00:00:01 |
| 2 | BITMAP CONVERSION TO ROWIDS| | 1500K| 10M| 132 (1)| 00:00:01 |
| 3 | BITMAP INDEX FULL SCAN | BLUB | | | | |
Statistiken
1 recursive calls
0 db block gets
1126 consistent gets
0 physical reads
0 redo size
378682 bytes sent via SQL*Net to client
11520 bytes received via SQL*Net from client
1002 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
15002 rows processed
-- the composite index
create index blub on tasks(START_DATE, 0);
SELECT START_DATE
, COUNT(START_DATE)
FROM TASKS
GROUP BY START_DATE
ORDER BY START_DATE;
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
| 0 | SELECT STATEMENT | | 15001 | 102K| 2400 (15)| 00:00:12 |
| 1 | SORT GROUP BY | | 15001 | 102K| 2400 (15)| 00:00:12 |
| 2 | INDEX FAST FULL SCAN| BLUB | 1500K| 10M| 2095 (3)| 00:00:11 |
Statistiken
0 recursive calls
0 db block gets
4120 consistent gets
0 physical reads
0 redo size
378682 bytes sent via SQL*Net to client
11520 bytes received via SQL*Net from client
1002 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
15002 rows processedRegards
Martin Preiss -
A Few Questions About Parallels 5 and Windows 7
Hey guys!
So, I am taking a computer programming class that requires Windows, just so we can all be on the same screen at the same time (lame, I know). Anywho, I have a Mac, so my teacher provided me with a log-in to MSDNAA, to get Windows 7 for free, and he gave me Parallels 5.0 for free as well. (go community colleges!)
A few questions:
1- I am running Snow Leopard right now on my 2009 MacBook Pro, and my processor is the 64-bit processor, but after running the command uname -a I discovered that my Kernal is loading in the 32-bit mode. Is this normal? Shouldn't it be running in 64-bit mode?
2- Which Windows 7 should I get, the x64 or x86 (32-bit or 64-bit)?
Thanks for any/all help!
Model Name: MacBook Pro
Model Identifier: MacBookPro5,4
Processor Name: Intel Core 2 Duo
Processor Speed: 2.53 GHz
Number Of Processors: 1
Total Number Of Cores: 2
L2 Cache: 3 MB
Memory: 4 GB
Chipset Model: NVIDIA GeForce 9400M
Type: GPU
Bus: PCI
VRAM (Total): 256 MB
Terminal reply after command:
local 10.4.0 Darwin Kernel Version 10.4.0: Fri Apr 23 18:28:53 PDT 2010; root:xnu-1504.7.4~1/RELEASE_I386 i386
the i386 means it's in 32-bit mode, right?Try the Windows forum area - http://discussions.apple.com/forum.jspa?forumID=687
-
Question about Parallels using Bootcamp partition
I was about to install XP on bootcamp but i would mainly be using parallels to run XP from the bootcamp partition. What i was wondering is that if i was to make any changes to xp like install a software and create a new folder through parallels, would it reflect on the actual bootcamp partition when i load into the actual XP partition using bootcamp? Basically, do the changes made to windows using parallels carry over to the bootcamp partition given you are using the bootcamp partition on parallels?
ThanksThanks a lot. While i am at it, if i use the bootcamp partition with Parallels, i cannot suspend the VM right?
Parallels will behave the same, whether it has its own copy of Windows, or is using that on the Boot Camp partition.
The only difference is backup. If you use Parallels with Windows in a disk image file (Preferably a sparse bundle), Windows can be backed up with Time Machine. If Windows is on its own partition, Time Machine can't handle it. -
Question about Resultant Query
Documentary Table
Doc_No Title Ranking YearReleased
1 Hoop Dreams A 1994
2 The Thin Blue Line B 2000
3 Surprise Me A 2004
4 Trouble the Water C 2008
5 Dark Days B 2000
6 Paris of Burning B 1991
7 Man on Wire A 2008
Please tell What will be the resultant relation that will be produced by the following SQL SELECT statement.
Query:
SELECT Doc_No, Title, YearReleased
FROM
Documentary
WHERE YearReleased <>‘2000’;
Please replyGhulam Yassen wrote:
I am asking about the relation after its output.
SELECT Doc_No, Title, YearReleased
FROM
Documentary
WHERE YearReleased ‘2000’;What do you mean with relation?
If you mean the output here you are:
WITH documentary AS
SELECT 1 Doc_no, 'Hoop Dreams' Title, 'A' Ranking, 1994 YearReleased FROM DUAL UNION
SELECT 2 Doc_no, 'The Thin Blue Line' Title, 'B' Ranking, 2000 YearReleased FROM DUAL UNION
SELECT 3 Doc_no, 'Surprise Me' Title, 'A' Ranking, 2004 YearReleased FROM DUAL UNION
SELECT 4 Doc_no, 'Trouble the Water' Title, 'C' Ranking, 2008 YearReleased FROM DUAL UNION
SELECT 5 Doc_no, 'Dark Days' Title, 'B' Ranking, 2000 YearReleased FROM DUAL UNION
SELECT 6 Doc_no, 'Paris of Burning' Title, 'B' Ranking, 1991 YearReleased FROM DUAL UNION
SELECT 7 Doc_no, 'Man on Wire' Title, 'A' Ranking, 2008 YearReleased FROM DUAL
SELECT doc_no, title, yearreleased
FROM documentary
WHERE yearreleased {noformat}<{noformat}{noformat}>{noformat} '2000';
DOC_NO TITLE YEARRELEASED
1 Hoop Dreams 1994
3 Surprise Me 2004
4 Trouble the Water 2008
6 Paris of Burning 1991
7 Man on Wire 2008
{code}
Regards.
Al
Edited by: Alberto Faenza on Jun 7, 2012 12:54 PM
Modified to show lt and gt signs -
Hi all,
I have created a rolequery and i made it as PUBLIC.
But when i go to fieldmapping and then selected the rolename radio button and when i tried to see the list in drop down box, i could not see my rolequery.
Did i miss anything?
Let me know if you do not understand the question.
Please help me.
Thank you.Hi,
In order to be able to select the rolequery in the dropdownlist, you have to create a new role. Preferably you copy an existing role, based on another existing rolwquery and name it like your rolequery.
When the role is created, add your role query to the role on the workflow routing page of the role. When this is done the role query will appear in the drop down list.
Kind regards,
Joris Verdonschot -
Question about different query results with wildcard
Hi, I'm working with a third party app on SQL Server 2000, and from what I can gather, programmed in C# & VisualFoxPro.
When we search with
Note contains 94949
we get 571 results, when we search with
Note contains 94949*
we get 575 results.
There should be at least a hundred different entries that start with "94949-1" so I expected the query with the wildcard to return something like 680 results, not an additional four rows.
Searching with
Note contains 94949-1*
got 483 results
Note contains 94949-10*
got 0 results
Could someone explain or point me to more documentation on the difference results we get?
ThanksHi Arnie, thanks for your response.
My situation is more basic than what appears in your example. Unfortunately, I am working solely through this application so I do not have access to the SQL to test out what you supplied, so my question is more of a need for an explanation of the search results. I do find lots of references to LIKE, Wildcards and pattern matching, but I don't find a way to explain to the users the best and most complete way to search. They mostly need a "this will always get us the results without missing anything" search technique and then to be able to select from a smaller group. I guess I need a basic course in understanding search results: how to get different ones and what they mean.
Using
Note contains 94949-'%'
returned one more result than when using an asterisk. I don't understand this difference.
Note contains 94949-'%1' or
Note contains 94949-'1%'
brings nothing nor does not using quotes. But there are hundreds of records which have the string starting with 94949-1 and a varying number of characters after that.
?Does the dash read not as a character in the string but as an expression?
When I use WITHIN 3 characters, I get too few results (eight). If I use AND, I get text unrelated to the account number I am looking for.
Again when I tried to narrow the search by adding one digit to the string to be matched, I did not get any results, but 500 results from the more general search is too much to scan by opening individual records.
Thanks for pondering this with me. -
Question about WebLogic query syntax
Hello,
I am using WebLogic Server 6.0, which came as part of the WebGain
Studio SE 4.5 development kit. My question regards the Web Logic query
syntax, which I have not yet mastered.
I am trying to create a finder method that takes a single argument
of type "char" and finds all matching fields of the column "keyword" in
which the argument is the first letter of keyword. That is, if I were
only looking for fields beginning with the letter "M", I'd use:
(like keyword 'M%')
However, I'm looking for all fields beginning with the first letter
defined by the first argument. Sadly, this syntax:
(like keyword '$0%')
doesn't seem to be working. Any suggestions on the correct syntax?
If this is not the right forum for this question, could someone
suggest an appropriate newsgroup?
Thanks, Dave962466 wrote:
Hi all,
I have an issue I need help with. I have created a script for an automated partition create on a monthly basis. It creates a monthly partition containing all dates within the respective month. The script is essentially this:
ALTER TABLE SCHEMA.TABLE
ADD PARTITION &&1
VALUES LESS THAN (to_number(to_char(to_date('&&2', 'DD-MON-YY'), 'YYYYMMDD')))
TABLESPACE LARGE_DATA94 COMPRESS;
I continually get this error message "ORA-14019: partition bound element must be one of: string, datetime or interval literal, number, or MAXVALUE"
The variable &&2 is passing in character data for the first of the month (E.G. '01-SEP-12'). &&1 passes character data for the month in MONYY (AUG12) I can run this query:
select
(to_number(to_char(to_date('&&2', 'DD-MON-YY'), 'YYYYMMDD')))
from dual;
With the output of 20120801. I cannot understand why I am able to run this partition create statement by hardcoding 20120901 but passing it in as a variable I receive the error. Note that I am not having problems with the &&1 variable. If anyone has any ideas please let me know. Thanks!I don't understand why you are taking a string, converting it to a date, then converting the date BACK TO a string ... then trying to convert that string to a number.
What is the data type of the partitioning key? It appears that it is a NUMBER, but actually represents a DATE. If so, that is a fundamentally flawed design. -
the documentation says that PARALLEL hint can by used in select statements like PARALLEL(emp 8). I have also seen syntaxes like PARALLEL, PARALLEL 8, PARALLEL(8), and in each case the steps in execution plans were the same but the cost assigned was a bit different between each version (but significantly lower than for a query without this hint).
how does oracle interpret above three syntaxes, ie. /*+ PARALLEL */, /*+ PARALLEL 8 */, /*+ PARALLEL(8) */?
are there any parts in those syntaxes that are ignored or incorporate default behaviour?
what syntax is safe to use when one doesn't want to specify the table being parallelized, ie. use syntax like PARALLEL(emp 8)?
thank you
Edited by: 943276 on 2012-07-25 19:17943276 wrote:
the documentation says that PARALLEL hint can by used like PARALLEL(emp 8). I have also seen syntaxes like PARALLEL, PARALLEL 8, PARALLEL(8), and in each case the steps in execution plans were the same but the cost assigned was different between each version (and significantly lower than for a query without this hint).
how does oracle interpret above three syntaxes, ie. /*+ PARALLEL */, /*+ PARALLEL 8 */, /*+ PARALLEL(8) */?
are there any parts in those syntaxes that are ignored or incorporate default behaviour?
what syntax is safe to use when one doesn't want to specify the table being parallelized, ie. use syntax like PARALLEL(emp 8)?
thank you
Edited by: 943276 on 2012-07-25 19:17You are NOT qualified to use any HINT.
Oracle recommends against using any HINT. -
Question about parallels desktop
Hello!
I have a little problem with my Parallels Desktop. I have installed a windows xp operating system with parallels desktop, but the virtual video card's memory is only 256Mb. In Mac OS X i have 2 video cards. One with 256Mb memory and one with 512Mb memory. Maybe can you help me how to solve this problem? Thanks in advance!:)Hi Tojas;
I would certainly hope that if you told the folks at Parallels that you were interested enough in their code to try a demo but that you were having problems with it that they would help you.
I would start at this site
http://forum.parallels.com/forumdisplay.php?f=58
Allan
Message was edited by: Allan Eckert -
Hi,
i want to insert the result of query to a table,
all the fields are corresponding except one field.
is there a way to this or i need to do
insert ( f1 , f2 ....)
i did this way
INTO CORRESPONDING FIELDS OF TABLE It_Return_Date
but i have another field which i want to insert to table.
is it possible??
10Xlets say you are getting values from a table
select & into X_TAB1 from TAB1.
data : x_tab2 type table2.
x_tab2-newfield = 'new value'. "Populate that new field value
move corresponding X_TAB1 TO X_TAB2. "all other will copy here
then use X_TAB2 to insert to your TABLE2.
INSERT TABLE2 FROM X_TAB2.
it would be a good practice to use same structure of the table while you use Insert/modify etc.. commands.
else you can declare TABLES ZTABLE2.
MOVE CORRESPONDING ZTABLE1 TO ZTABLE2.
ZTABLE2-NEW_FIELD = <NEW VALUE>.
then use INSERT.
INSERT ZTABLE2.
Regards
srikanth
Message was edited by: Srikanth Kidambi -
Question about infoset query with code
Hi experts.
I always appreciate your help. Today, I have some problem. I made infoset query using SQ01 and got some data. Those data include WBS no but some of them didn't. So I tried to put some code into the infoset query that any data has not WBS has some value, for example 'X', in WBS no field.
I needed to know what field or table I should control to display value 'X' in WBS no field in case that WBS field has no value.
I tried to put code at "END-OF-SELECTION(before display)" and control %dtab. But I met error message there is no such table %dtab. So, I define %dtab then during runtime, I also met error message %dtab has already been defiend.
How can I do? Is it impossible thing?? I'm wating for your answer.
regards.Hi, if you just want to fill the field if it is blank, you can do this in the Record Processing Section.
I'm not familar with the WBS number, so for this example, I will fill the BISMT in MARA during the "Recording Processing" Section.
if mara-bismt is initial.
mara-bismt = 'X'.
endif.
Regards,
Rich Heilman
Maybe you are looking for
-
Web service proxy client with client cert cause SSLSessionNotFoundErr
Hi, I tried to run web service proxy client with certification from JDeveloper 10.1.3.0.4 to call PKI enabled web service got folllowing error: WARNING: Unable to connect to URL: due to java.security.PrivilegedActionException: javax.xml.soap.SOAPExce
-
Include ZXM06U02 in EXIT_SAPLEINM_002
Hi gurus, i've tried to add my z-segment ZE1EDP01 into int_edidd like it is described in the following link: http://help.sap.com/saphelp_dimp50/helpdata/DE/dc/6b7d7c43d711d1893e0000e8323c4f/content.htm The z-segment will be added, but the idoc went t
-
Just got my new MacbookPro 17in. I'm having problems when the energy saver kicks in and the screen starts to go dim, when I click any key to wake it back up the entire screen goes 90% black. Then after trying to get it to stop doing that the screen j
-
I can not get my phone to sign in to my wi fi. I used to, then I lost power to the internet and now it won't. All other devices will.
-
Hi, From what tables are taken data when is processed the code? get objec. I suppose, reading other posts, that main table is hrp1000. Is it right?