Parallel process in select query

Hi Experts
its production
stats are good
indexes are good
issue :
a select query is taking time 20min , the indexes exists on the where clause columns, I tried the query using parallel hint it gave the results iin 20 sec.
the DOP of the table set to 10, but when I run the query it picks old plan
constraints : I cannot modify the query, it needs to be tuned without modifying it
if DOP is >1 then select is expected to use parallellism in explain plan..is my understanding correct?
if I am wrong please advice how can i force the query to use the parallellism
Please advice and guide asap
THanks in advance
Ajay Kumar

user513478 wrote:
Hi Experts
its production
stats are good
indexes are good
issue :
a select query is taking time 20min , the indexes exists on the where clause columns, I tried the query using parallel hint it gave the results iin 20 sec.
the DOP of the table set to 10, but when I run the query it picks old plan
constraints : I cannot modify the query, it needs to be tuned without modifying itWhat version of Oracle?
Tuning SQL without modifying the SQL can be very hard but you may have options.
Could you use a materialized view with automatic query rewrite?
Can you use a SQL profile to affect performance?
Have you looked very+ carefully at the indexes to see if other indexes might be more efficient?
20 minutes to 20 seconds sounds too good to be true. Are you sure you aren't reading cached data and that the 2nd, faster run really is running in parallel?

Similar Messages

How to limit number of parallel processes for a query???

Hi,
I have set table parallelism to degree (4 in my case) and when i run a query on that table i see on v$session that this query is using 8 parallel processes.
Why my query is using all this processes?? Can i limit this number?? I think this will cause poor performance if all my parallel processes stay BUSY on v$pq_slave.
1 1 P000 BUSY 22 0 8 0 2051 10 2 15 0 79179 76884
2 1 P001 BUSY 22 0 8 0 2054 10 2 15 0 81905 77443
3 1 P004 BUSY 2 0 1592 0 0 0 0 1592 0 1039 3
4 1 P005 BUSY 2 0 1592 0 0 0 0 1592 0 1038 4
5 1 PZ99 BUSY 533 0 0 0 1 3 0 0 0 1071 1107
6 2 P000 BUSY 14 0 8 0 2053 10 3 15 1 53014 73297
7 2 P001 BUSY 14 0 8 0 2048 10 2 15 1 51266 73318
8 2 P002 BUSY 14 0 8 0 2052 10 2 15 2 51043 73271
9 2 P003 BUSY 14 0 8 0 2053 9 2 15 2 49417 73327
10 2 P004 BUSY 13 0 8 0 2055 9 2 15 2 68428 12468
11 2 P005 BUSY 13 0 8 0 2059 10 2 15 1 69968 12473
12 2 PZ99 BUSY 461 0 0 0 1 3 0 0 0 921 936
Tks,
Paulo.

select /*+ PARALLEL(a,4) */ ...... from owner.table a;
Or
ALTER SESSION FORCE PARALLEL DML PARALLEL <degree>
(But I am not sure whether this will affect the degree of select also ?)

In oracle rac, If user query a select query and in processing data is fetched but in the duration of fetching the particular node is evicted then how failover to another node internally?

In oracle rac, If user query a select query and in processing data is fetched but in the duration of fetching the particular node is evicted then how failover to another node internally?

The query is re-issued as a flashback query and the client process can continue to fetch from the cursor. This is described in the Net Services Administrators Guide, the section on Transparent Application Failover.

Reg: SQL select Query in BPEL process flow


Hi,
I am suppose to execute a SQL select query (in BPEL Process flow) as mention below in JDeveloper using Database adapter.


SELECT LENGTH, WIDTH, HEIGHT, WEIGHT,


LENGTH*WIDTH* HEIGHT AS ITEM_CUBE


FROM CUBE


WHERE ITEM= <xyz>


AND OBJECT= (SELECT CASE_NAME FROM CUBE_SUPPLIER WHERE ITEM=<xyz> AND SUPP_IND = ‘Y')
Now my question is:
1. What does this "*" refer to in the query and how can I retrieve the value of LENGTH*WIDTH* HEIGHT from the query where LENGTH,WIDTH and HEIGHT are the individual field in the table.
2.What does this " AS" refer to? If " ITEM_CUBE " is the alies for the table name "ITEM" to retrieve the value, then query shoud be evaluated as


SELECT LENGTH, WIDTH, HEIGHT, WEIGHT,


LENGTH*WIDTH* HEIGHT AS ITEM_CUBE


FROM CUBE


WHERE ITEM_CUBE.ITEM= <xyz>


AND ITEM_CUBE.OBJECT= (SELECT CASE_NAME FROM CUBE_SUPPLIER WHERE ITEM=<xyz> AND SUPP_IND = ‘Y')
Is my assumption correct?
Please suggest asap.
Thanks...




Hi
Thank for your reply!
I have a nested select query which performs on two different table as shown below:

SELECT LENGTH, WIDTH, HEIGHT, WEIGHT,


LENGTH*WIDTH* HEIGHT AS ITEM_CUBE


FROM CUBE


WHERE ITEM= <abc>


AND OBJECT= (SELECT NAME FROM SUPPLIER WHERE ITEM=<Item> AND SUPP_IND = ‘Y')
I am using DB adapter of Oracle JDeveloper in BPEL process flow, where I can able to select only one master table in DB adapter say SUPPLIER and its attributes at a time.But as per my requirment I need to select both the table (CUBE and SUPPLIER) in a single adapter to execute my query.
It can be achievable by using two DB adapter , One to execute the nested query and another to execute the main qyery considering value of nested query as a parameter.But I want to achieve it by using a single one.
Am I correct with my concept?
Please suggest how to get it ?

Edited by: user10259700 on Oct 23, 2008 12:17 AM

CV04N takes long time to process select query on DRAT table

Hello Team,
While using CV04N to display DIR's, it takes long time to process select query on DRAT table. This query includes all the key fields. Any idea as to how to analyse this?
Thanks and best regards,
Bobby
Moderator message: please read the sticky threads of this forum, there is a lot of information on what you can do.
Edited by: Thomas Zloch on Feb 24, 2012

Be aware that XP takes approx 1gb of your RAM leaving you with 1gb for whatever else is running. MS Outlook is also a memory hog.
To check Virtual Memory Settings:
Control Panel -> System
System Properties -> Advanced Tab -> Performance Settings
Performance Options -> Adavanced Tab - Virtual Memory section
Virtual Memory -
what are
* Initial Size
* Maximum Size
In a presentation at one of the Hyperion conferences years ago, Mark Ostroff suggested that the initial be set to the same as Max. (Max is typically 2x physical RAM)
These changes may provide some improvement.

Disjoint Selections in Parallel Processing Profile

Hello All,
I wanted to use the parallel processing profile for my DP background jobs.
When I referred the 'Note 853142 - Consulting: Parallel processing profile for DP batch runs' it says use this parallel processing using disjoint selections (external parallel processing). Can you please explain what is the meaning of disjoint selections.
Thanks,
Siva.

Hi Siva,
With quite large background runs, the background work process can experience
memory problems when the characteristics combinations are being imported.
Therefore, an additional external' parallel processing (using several background jobs
and disjoint selections) may be necessary. Disjoint selections has no link to its
selections made.
Regards
R. Senthil Mareeswaran.

Strange responsetime with query triggering parallel processing

Hi,
Are there anyone who can run this SQL below on a RAC 11g R2, and give me feedback on how long time it took?
In my environment it takes more than a minute, and I suspect there is a problem with the database and parallel processing.
If I collect the values from the inner select and make a list of comma-separated values, it goes fast.
Replace SCHEMA_NAME with an existing SCHEMA_NAME on you database before running.
SELECT sa.*
FROM gv$sqlarea sa
WHERE sa.sql_id in (SELECT sql_id from gv$sqlarea where parsing_schema_name like '%SCHEMA_NAME%' and rownum < 100)
Thanks!

You are using the hint a bit incorrectly.
http://download.oracle.com/docs/cd/B19306_01/server.102/b14200/sql_elements006.htm#SQLRF50805
<quote>
The NO_PARALLEL hint overrides a PARALLEL parameter in the DDL that created or altered the table. For example:
SELECT /*+ NO_PARALLEL(hr_emp) */ last_name
FROM employees hr_emp;
<quote>
Check Metalink note 267330.1 as well.

Parallel processing query

Hello all,
I know parallel processing improves data performance when loading from source system. but how is parallel processing done in BW? how can we set up the number of parallel process on BW system? And does it have any impact on any other processes?
Thanks

S N
Welcome to SDN
please see this thread
Re: Parallel data loading
useful links
http://help.sap.com/saphelp_nw04/helpdata/en/c4/3a7f1f505211d189550000e829fbbd/frameset.htm
http://help.sap.com/saphelp_nw04/helpdata/en/fa/096ff6543b11d1898e0000e8322d00/frameset.htm
Hope this helps
Thnaks
sat

Getting deadlock detected while waiting for resource error for select Query.....

Hi all,
i am getting a below error whenever executing the below select query....
some times it will show dead lock detected while waiting for resource and terminated...
some times it executes and gives result..
but all the time it writes an alert to alert log
Plesae suggest how to resolve the issue..........
Thanks in advance
Env: Linux / Oracle 11.2.0.3.3
Error from alert log:
Errors in file /u01/oracle/oracle/diag/rdbms/bdrdb/bdrdb/trace/bdrdb_p017_6076.trc:
ORA-00060: deadlock detected while waiting for resource
ORA-10387: parallel query server interrupt (normal)
Trace file info... bdrdb_p017_6076.trc:
Trace file /u01/oracle/oracle/diag/rdbms/bdrdb/bdrdb/trace/bdrdb_p017_6076.trc
Oracle Database 11g Enterprise Edition Release 11.2.0.3.0 - Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options
ORACLE_HOME = /u01/oracle/oracle/product/11.2.0/dbhome_1
System name: Linux
Node name: bdrdb.cteplindia.com
Release: 2.6.18-308.el5PAE
Version: #1 SMP Fri Jan 27 17:40:09 EST 2012
Machine: i686
Instance name: bdrdb
Redo thread mounted by this instance: 1
Oracle process number: 92
Unix process pid: 6076, image: [email protected] (P017)
*** 2013-11-04 23:18:57.915
*** SESSION ID:(423.59970) 2013-11-04 23:18:57.915
*** CLIENT ID:() 2013-11-04 23:18:57.915
*** SERVICE NAME:(bdrdb) 2013-11-04 23:18:57.915
*** MODULE NAME:() 2013-11-04 23:18:57.915
*** ACTION NAME:() 2013-11-04 23:18:57.915
*** 2013-11-04 23:18:57.915
DEADLOCK DETECTED ( ORA-00060 )
[Transaction Deadlock]
Deadlock graph:
 ---------Blocker(s)-------- ---------Waiter(s)---------
Resource Name process session holds waits process session holds waits
PS-00000001-00000011 92 423 S 33 128 S X
BF-2ed08c01-00000000 33 128 S 92 423 S X
session 423: DID 0001-005C-00081126 session 128: DID 0001-0021-00067D23
session 128: DID 0001-0021-00067D23 session 423: DID 0001-005C-00081126
DEADLOCK DETECTED ( ORA-00060 )
[Transaction Deadlock]
Deadlock graph:
 ---------Blocker(s)-------- ---------Waiter(s)---------
Resource Name process session holds waits process session holds waits
PS-00000001-00000011 92 423 S 33 128 S X
BF-2ed08c01-00000000 33 128 S 92 423 S X
session 423: DID 0001-005C-00081126 session 128: DID 0001-0021-00067D23
session 128: DID 0001-0021-00067D23 session 423: DID 0001-005C-00081126
Rows waited on:
Session 423: no row
Session 128: obj - rowid = 00021DC1 - AAAh3BAAVAAAQL/AAA
(dictionary objn - 138689, file - 21, block - 66303, slot - 0)
----- Information for the OTHER waiting sessions -----
Session 128:
sid: 128 ser: 46176 audsid: 1836857 user: 102/DBLOCAL
 flags: (0x8000041) USR/- flags_idl: (0x1) BSY/-/-/-/-/-
 flags2: (0x40009) -/-/INC
pid: 33 O/S info: user: oracle, term: UNKNOWN, ospid: 31611
 image: [email protected]
client details:
 O/S info: user: masked, term: masked, ospid: 5924:568
 machine: masked program: Toad.exe
 application name: TOAD background query session, hash value=526966934
current SQL:
 application name: TOAD background query session, hash value=526966934
current SQL:
SELECT DISTINCT B_FP_TEST.TEST_ID
FROM B_FP_TEST,
 B_USER_INFO,
 J_FP_INVESTIGATOR,
 L_TEST_STATUS,
 L_ATMS_TEST_TYPE,
 j_op_test_anml
WHERE B_FP_TEST.TEST_ID = J_FP_INVESTIGATOR.TEST_ID
 AND B_FP_TEST.TEST_TYPE_ID = L_ATMS_TEST_TYPE.ATMS_TEST_TYPE_ID
 AND B_USER_INFO.B_USER_INFO_ID = J_FP_INVESTIGATOR.INVESTIGATOR_ID
 AND B_FP_TEST.STATUS_ID = L_TEST_STATUS.STATUS_ID
 AND B_FP_TEST.IS_DELETED = :"SYS_B_00"
 AND B_FP_TEST.TEST_NUM NOT IN (:"SYS_B_01", :"SYS_B_02", :"SYS_B_03")
 AND L_ATMS_TEST_TYPE.IS_DELETED = :"SYS_B_04"
 AND J_FP_INVESTIGATOR.is_pi = :"SYS_B_05"
 AND L_TEST_STATUS.STATUS IN (:"SYS_B_06", :"SYS_B_07", :"SYS_B_08")
 AND j_op_test_anml.test_id = B_FP_TEST.TEST_ID
----- End of information for the OTHER waiting sessions -----
*** 2013-11-04 23:18:57.916
dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x0, level=3, mask=0x0)
----- Error Stack Dump -----
ORA-00060: deadlock detected while waiting for resource
ORA-10387: parallel query server interrupt (normal)
----- SQL Statement (None) -----
Current SQL information unavailable - no cursor.
----- Call Stack Trace -----
calling call entry argument values in hex
location type point (? means dubious value)
More......
Query:
SELECT DISTINCT B_FP_TEST.TEST_ID
FROM B_FP_TEST,
 B_USER_INFO,
 J_FP_INVESTIGATOR,
 L_TEST_STATUS,
 L_ATMS_TEST_TYPE,
 j_op_test_anml
WHERE B_FP_TEST.TEST_ID = J_FP_INVESTIGATOR.TEST_ID
 AND B_FP_TEST.TEST_TYPE_ID = L_ATMS_TEST_TYPE.ATMS_TEST_TYPE_ID
 AND B_USER_INFO.B_USER_INFO_ID = J_FP_INVESTIGATOR.INVESTIGATOR_ID
 AND B_FP_TEST.STATUS_ID = L_TEST_STATUS.STATUS_ID
 AND B_FP_TEST.IS_DELETED = 0
 AND B_FP_TEST.TEST_NUM NOT IN (1, 2, 99)
 AND L_ATMS_TEST_TYPE.IS_DELETED = 0
 AND J_FP_INVESTIGATOR.is_pi = 1
 AND L_TEST_STATUS.STATUS IN ('Scheduled', 'In-Progress', 'Completed')
 AND j_op_test_anml.test_id = B_FP_TEST.TEST_ID
 AND ( (j_op_test_anml.end_date BETWEEN TO_DATE ('28-Oct-2013') - 1
 AND TO_DATE ('04-Nov-2013') + 1)
 OR (j_op_test_anml.start_date BETWEEN TO_DATE ('28-Oct-2013') - 1
 AND TO_DATE ('04-Nov-2013') + 1)
 OR (TO_DATE ('28-Oct-2013') BETWEEN j_op_test_anml.start_date
 AND j_op_test_anml.end_date)
 OR (TO_DATE ('04-Nov-2013') BETWEEN j_op_test_anml.start_date
 AND j_op_test_anml.end_date))
 AND L_ATMS_TEST_TYPE.IS_DELETED = 0
 AND B_FP_TEST.DATASOURCE_ID = 9
Query Exp plan:
Plan hash value: 3398228788
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | Pstart| Pstop | TQ |IN-OUT| PQ Distrib |
| 0 | SELECT STATEMENT | | 1501 | 102K| 1929 (1)| 00:00:24 | | | | | |
| 1 | HASH UNIQUE | | 1501 | 102K| 1929 (1)| 00:00:24 | | | | | |
| 2 | CONCATENATION | | | | | | | | | | |
| 3 | PX COORDINATOR | | | | | | | | | | |
| 4 | PX SEND QC (RANDOM) | :TQ30005 | 241 | 16870 | 800 (1)| 00:00:10 | | | Q3,05 | P->S | QC (RAND) |
|* 5 | HASH JOIN | | 241 | 16870 | 800 (1)| 00:00:10 | | | Q3,05 | PCWP | |
| 6 | PX RECEIVE | | 246 | 15990 | 797 (1)| 00:00:10 | | | Q3,05 | PCWP | |
| 7 | PX SEND HASH | :TQ30004 | 246 | 15990 | 797 (1)| 00:00:10 | | | Q3,04 | P->P | HASH |
|* 8 | HASH JOIN | | 246 | 15990 | 797 (1)| 00:00:10 | | | Q3,04 | PCWP | |
| 9 | PX RECEIVE | | 573 | 29223 | 793 (1)| 00:00:10 | | | Q3,04 | PCWP | |
| 10 | PX SEND HASH | :TQ30003 | 573 | 29223 | 793 (1)| 00:00:10 | | | Q3,03 | P->P | HASH |
|* 11 | HASH JOIN | | 573 | 29223 | 793 (1)| 00:00:10 | | | Q3,03 | PCWP | |
| 12 | BUFFER SORT | | | | | | | | Q3,03 | PCWC | |
| 13 | PX RECEIVE | | | | | | | | Q3,03 | PCWP | |
| 14 | PX SEND BROADCAST | :TQ30000 | | | | | | | | S->P | BROADCAST |
| 15 | NESTED LOOPS | | | | | | | | | | |
| 16 | NESTED LOOPS | | 485 | 20855 | 781 (0)| 00:00:10 | | | | | |
| 17 | TABLE ACCESS BY GLOBAL INDEX ROWID| J_OP_TEST_ANML | 485 | 10185 | 296 (0)| 00:00:04 | ROWID | ROWID | | | |
|* 18 | INDEX RANGE SCAN | IDX$$_2D190001 | 485 | | 4 (0)| 00:00:01 | | | | | |
|* 19 | INDEX UNIQUE SCAN | FT_TEST_ID_PK | 1 | | 0 (0)| 00:00:01 | | | | | |
|* 20 | TABLE ACCESS BY GLOBAL INDEX ROWID | B_FP_TEST | 1 | 22 | 1 (0)| 00:00:01 | ROWID | ROWID | | | |
| 21 | PX BLOCK ITERATOR | | 70382 | 549K| 11 (0)| 00:00:01 | | | Q3,03 | PCWC | |
|* 22 | TABLE ACCESS FULL | J_FP_INVESTIGATOR | 70382 | 549K| 11 (0)| 00:00:01 | | | Q3,03 | PCWP | |
| 23 | BUFFER SORT | | | | | | | | Q3,04 | PCWC | |
| 24 | PX RECEIVE | | 3 | 42 | 3 (0)| 00:00:01 | | | Q3,04 | PCWP | |
| 25 | PX SEND HASH | :TQ30001 | 3 | 42 | 3 (0)| 00:00:01 | | | | S->P | HASH |
|* 26 | TABLE ACCESS FULL | L_TEST_STATUS | 3 | 42 | 3 (0)| 00:00:01 | | | | | |
| 27 | BUFFER SORT | | | | | | | | Q3,05 | PCWC | |
| 28 | PX RECEIVE | | 30 | 150 | 3 (0)| 00:00:01 | | | Q3,05 | PCWP | |
| 29 | PX SEND HASH | :TQ30002 | 30 | 150 | 3 (0)| 00:00:01 | | | | S->P | HASH |
|* 30 | TABLE ACCESS FULL | L_ATMS_TEST_TYPE | 30 | 150 | 3 (0)| 00:00:01 | | | | | |
| 31 | NESTED LOOPS | | | | | | | | | | |
| 32 | NESTED LOOPS | | 3 | 210 | 329 (1)| 00:00:04 | | | | | |
| 33 | NESTED LOOPS | | 3 | 195 | 329 (1)| 00:00:04 | | | | | |
|* 34 | HASH JOIN | | 2 | 114 | 325 (1)| 00:00:04 | | | | | |
| 35 | NESTED LOOPS | | | | | | | | | | |
| 36 | NESTED LOOPS | | 6 | 258 | 322 (1)| 00:00:04 | | | | | |
| 37 | PARTITION RANGE SINGLE | | 6 | 126 | 316 (1)| 00:00:04 | 7 | 7 | | | |
|* 38 | TABLE ACCESS FULL | J_OP_TEST_ANML | 6 | 126 | 316 (1)| 00:00:04 | 7 | 7 | | | |
|* 39 | INDEX UNIQUE SCAN | FT_TEST_ID_PK | 1 | | 0 (0)| 00:00:01 | | | | | |
|* 40 | TABLE ACCESS BY GLOBAL INDEX ROWID | B_FP_TEST | 1 | 22 | 1 (0)| 00:00:01 | ROWID | ROWID | | | |
|* 41 | TABLE ACCESS FULL | L_TEST_STATUS | 3 | 42 | 3 (0)| 00:00:01 | | | | | |
|* 42 | TABLE ACCESS BY INDEX ROWID | J_FP_INVESTIGATOR | 1 | 8 | 2 (0)| 00:00:01 | | | | | |
|* 43 | INDEX RANGE SCAN | FI_TEST_ID_PK | 1 | | 1 (0)| 00:00:01 | | | | | |
|* 44 | INDEX UNIQUE SCAN | L_ATMS_TEST_TYPE_PK | 1 | | 0 (0)| 00:00:01 | | | | | |
|* 45 | TABLE ACCESS BY INDEX ROWID | L_ATMS_TEST_TYPE | 1 | 5 | 1 (0)| 00:00:01 | | | | | |
| 46 | PX COORDINATOR | | | | | | | | | | |
| 47 | PX SEND QC (RANDOM) | :TQ20003 | | | | | | | Q2,03 | P->S | QC (RAND) |
| 48 | NESTED LOOPS | | | | | | | | Q2,03 | PCWP | |
| 49 | NESTED LOOPS | | 33 | 2310 | 399 (2)| 00:00:05 | | | Q2,03 | PCWP | |
|* 50 | HASH JOIN | | 33 | 2145 | 397 (2)| 00:00:05 | | | Q2,03 | PCWP | |
| 51 | PX RECEIVE | | 78 | 3978 | 393 (1)| 00:00:05 | | | Q2,03 | PCWP | |
| 52 | PX SEND HASH | :TQ20002 | 78 | 3978 | 393 (1)| 00:00:05 | | | Q2,02 | P->P | HASH |
|* 53 | HASH JOIN | | 78 | 3978 | 393 (1)| 00:00:05 | | | Q2,02 | PCWP | |
| 54 | BUFFER SORT | | | | | | | | Q2,02 | PCWC | |
| 55 | PX RECEIVE | | | | | | | | Q2,02 | PCWP | |
| 56 | PX SEND BROADCAST | :TQ20000 | | | | | | | | S->P | BROADCAST |
| 57 | NESTED LOOPS | | | | | | | | | | |
| 58 | NESTED LOOPS | | 66 | 2838 | 382 (1)| 00:00:05 | | | | | |
| 59 | PARTITION RANGE SINGLE | | 66 | 1386 | 316 (1)| 00:00:04 | 7 | 7 | | | |
|* 60 | TABLE ACCESS FULL | J_OP_TEST_ANML | 66 | 1386 | 316 (1)| 00:00:04 | 7 | 7 | | | |
|* 61 | INDEX UNIQUE SCAN | FT_TEST_ID_PK | 1 | | 0 (0)| 00:00:01 | | | | | |
|* 62 | TABLE ACCESS BY GLOBAL INDEX ROWID | B_FP_TEST | 1 | 22 | 1 (0)| 00:00:01 | ROWID | ROWID | | | |
| 63 | PX BLOCK ITERATOR | | 70382 | 549K| 11 (0)| 00:00:01 | | | Q2,02 | PCWC | |
|* 64 | TABLE ACCESS FULL | J_FP_INVESTIGATOR | 70382 | 549K| 11 (0)| 00:00:01 | | | Q2,02 | PCWP | |
| 65 | BUFFER SORT | | | | | | | | Q2,03 | PCWC | |
| 66 | PX RECEIVE | | 3 | 42 | 3 (0)| 00:00:01 | | | Q2,03 | PCWP | |
| 67 | PX SEND HASH | :TQ20001 | 3 | 42 | 3 (0)| 00:00:01 | | | | S->P | HASH |
|* 68 | TABLE ACCESS FULL | L_TEST_STATUS | 3 | 42 | 3 (0)| 00:00:01 | | | | | |
|* 69 | INDEX UNIQUE SCAN | L_ATMS_TEST_TYPE_PK | 1 | | 0 (0)| 00:00:01 | | | Q2,03 | PCWP | |
|* 70 | TABLE ACCESS BY INDEX ROWID | L_ATMS_TEST_TYPE | 1 | 5 | 1 (0)| 00:00:01 | | | Q2,03 | PCWP | |
| 71 | PX COORDINATOR | | | | | | | | | | |
| 72 | PX SEND QC (RANDOM) | :TQ10003 | | | | | | | Q1,03 | P->S | QC (RAND) |
| 73 | NESTED LOOPS | | | | | | | | Q1,03 | PCWP | |
| 74 | NESTED LOOPS | | 33 | 2310 | 399 (2)| 00:00:05 | | | Q1,03 | PCWP | |
|* 75 | HASH JOIN | | 34 | 2210 | 397 (2)| 00:00:05 | | | Q1,03 | PCWP | |
| 76 | PX RECEIVE | | 78 | 3978 | 393 (1)| 00:00:05 | | | Q1,03 | PCWP | |
| 77 | PX SEND HASH | :TQ10002 | 78 | 3978 | 393 (1)| 00:00:05 | | | Q1,02 | P->P | HASH |
|* 78 | HASH JOIN | | 78 | 3978 | 393 (1)| 00:00:05 | | | Q1,02 | PCWP | |
| 79 | BUFFER SORT | | | | | | | | Q1,02 | PCWC | |
| 80 | PX RECEIVE | | | | | | | | Q1,02 | PCWP | |
| 81 | PX SEND BROADCAST | :TQ10000 | | | | | | | | S->P | BROADCAST |
| 82 | NESTED LOOPS | | | | | | | | | | |
| 83 | NESTED LOOPS | | 66 | 2838 | 382 (1)| 00:00:05 | | | | | |
| 84 | PARTITION RANGE SINGLE | | 66 | 1386 | 316 (1)| 00:00:04 | 7 | 7 | | | |
|* 85 | TABLE ACCESS FULL | J_OP_TEST_ANML | 66 | 1386 | 316 (1)| 00:00:04 | 7 | 7 | | | |
|* 86 | INDEX UNIQUE SCAN | FT_TEST_ID_PK | 1 | | 0 (0)| 00:00:01 | | | | | |
|* 87 | TABLE ACCESS BY GLOBAL INDEX ROWID | B_FP_TEST | 1 | 22 | 1 (0)| 00:00:01 | ROWID | ROWID | | | |
| 88 | PX BLOCK ITERATOR | | 70382 | 549K| 11 (0)| 00:00:01 | | | Q1,02 | PCWC | |
|* 89 | TABLE ACCESS FULL | J_FP_INVESTIGATOR | 70382 | 549K| 11 (0)| 00:00:01 | | | Q1,02 | PCWP | |
| 90 | BUFFER SORT | | | | | | | | Q1,03 | PCWC | |
| 91 | PX RECEIVE | | 3 | 42 | 3 (0)| 00:00:01 | | | Q1,03 | PCWP | |
| 92 | PX SEND HASH | :TQ10001 | 3 | 42 | 3 (0)| 00:00:01 | | | | S->P | HASH |
|* 93 | TABLE ACCESS FULL | L_TEST_STATUS | 3 | 42 | 3 (0)| 00:00:01 | | | | | |
|* 94 | INDEX UNIQUE SCAN | L_ATMS_TEST_TYPE_PK | 1 | | 0 (0)| 00:00:01 | | | Q1,03 | PCWP | |
|* 95 | TABLE ACCESS BY INDEX ROWID | L_ATMS_TEST_TYPE | 1 | 5 | 1 (0)| 00:00:01 | | | Q1,03 | PCWP | |
Predicate Information (identified by operation id):
 5 - access("B_FP_TEST"."TEST_TYPE_ID"="L_ATMS_TEST_TYPE"."ATMS_TEST_TYPE_ID")
 8 - access("B_FP_TEST"."STATUS_ID"="L_TEST_STATUS"."STATUS_ID")
11 - access("B_FP_TEST"."TEST_ID"="J_FP_INVESTIGATOR"."TEST_ID")
18 - access("J_OP_TEST_ANML"."START_DATE">=TO_DATE(' 2013-10-27 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND "J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-11-05
 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))
19 - access("J_OP_TEST_ANML"."TEST_ID"="B_FP_TEST"."TEST_ID")
20 - filter("B_FP_TEST"."DATASOURCE_ID"=9 AND "B_FP_TEST"."IS_DELETED"=0 AND "B_FP_TEST"."TEST_NUM"<>1 AND "B_FP_TEST"."TEST_NUM"<>2 AND
 "B_FP_TEST"."TEST_NUM"<>99)
22 - filter("J_FP_INVESTIGATOR"."IS_PI"=1)
26 - filter("L_TEST_STATUS"."STATUS"='Completed' OR "L_TEST_STATUS"."STATUS"='In-Progress' OR "L_TEST_STATUS"."STATUS"='Scheduled')
30 - filter("L_ATMS_TEST_TYPE"."IS_DELETED"=0)
34 - access("B_FP_TEST"."STATUS_ID"="L_TEST_STATUS"."STATUS_ID")
38 - filter("J_OP_TEST_ANML"."END_DATE">=TO_DATE(' 2013-10-27 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND "J_OP_TEST_ANML"."END_DATE"<=TO_DATE(' 2013-11-05
 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND (LNNVL("J_OP_TEST_ANML"."START_DATE">=TO_DATE(' 2013-10-27 00:00:00', 'syyyy-mm-dd hh24:mi:ss')) OR
 LNNVL("J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-11-05 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))))
39 - access("J_OP_TEST_ANML"."TEST_ID"="B_FP_TEST"."TEST_ID")
40 - filter("B_FP_TEST"."DATASOURCE_ID"=9 AND "B_FP_TEST"."IS_DELETED"=0 AND "B_FP_TEST"."TEST_NUM"<>1 AND "B_FP_TEST"."TEST_NUM"<>2 AND
 "B_FP_TEST"."TEST_NUM"<>99)
41 - filter("L_TEST_STATUS"."STATUS"='Completed' OR "L_TEST_STATUS"."STATUS"='In-Progress' OR "L_TEST_STATUS"."STATUS"='Scheduled')
42 - filter("J_FP_INVESTIGATOR"."IS_PI"=1)
43 - access("B_FP_TEST"."TEST_ID"="J_FP_INVESTIGATOR"."TEST_ID")
44 - access("B_FP_TEST"."TEST_TYPE_ID"="L_ATMS_TEST_TYPE"."ATMS_TEST_TYPE_ID")
45 - filter("L_ATMS_TEST_TYPE"."IS_DELETED"=0)
50 - access("B_FP_TEST"."STATUS_ID"="L_TEST_STATUS"."STATUS_ID")
53 - access("B_FP_TEST"."TEST_ID"="J_FP_INVESTIGATOR"."TEST_ID")
60 - filter("J_OP_TEST_ANML"."END_DATE">=TO_DATE(' 2013-11-04 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND "J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-11-04
 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND (LNNVL("J_OP_TEST_ANML"."END_DATE">=TO_DATE(' 2013-10-27 00:00:00', 'syyyy-mm-dd hh24:mi:ss')) OR
 LNNVL("J_OP_TEST_ANML"."END_DATE"<=TO_DATE(' 2013-11-05 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))) AND (LNNVL("J_OP_TEST_ANML"."START_DATE">=TO_DATE(' 2013-10-27
 00:00:00', 'syyyy-mm-dd hh24:mi:ss')) OR LNNVL("J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-11-05 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))))
61 - access("J_OP_TEST_ANML"."TEST_ID"="B_FP_TEST"."TEST_ID")
62 - filter("B_FP_TEST"."DATASOURCE_ID"=9 AND "B_FP_TEST"."IS_DELETED"=0 AND "B_FP_TEST"."TEST_NUM"<>1 AND "B_FP_TEST"."TEST_NUM"<>2 AND
 "B_FP_TEST"."TEST_NUM"<>99)
64 - filter("J_FP_INVESTIGATOR"."IS_PI"=1)
68 - filter("L_TEST_STATUS"."STATUS"='Completed' OR "L_TEST_STATUS"."STATUS"='In-Progress' OR "L_TEST_STATUS"."STATUS"='Scheduled')
69 - access("B_FP_TEST"."TEST_TYPE_ID"="L_ATMS_TEST_TYPE"."ATMS_TEST_TYPE_ID")
70 - filter("L_ATMS_TEST_TYPE"."IS_DELETED"=0)
75 - access("B_FP_TEST"."STATUS_ID"="L_TEST_STATUS"."STATUS_ID")
78 - access("B_FP_TEST"."TEST_ID"="J_FP_INVESTIGATOR"."TEST_ID")
85 - filter("J_OP_TEST_ANML"."END_DATE">=TO_DATE(' 2013-10-28 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND "J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-10-28
 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND (LNNVL("J_OP_TEST_ANML"."END_DATE">=TO_DATE(' 2013-11-04 00:00:00', 'syyyy-mm-dd hh24:mi:ss')) OR
 LNNVL("J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-11-04 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))) AND (LNNVL("J_OP_TEST_ANML"."END_DATE">=TO_DATE(' 2013-10-27
 00:00:00', 'syyyy-mm-dd hh24:mi:ss')) OR LNNVL("J_OP_TEST_ANML"."END_DATE"<=TO_DATE(' 2013-11-05 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))) AND
 (LNNVL("J_OP_TEST_ANML"."START_DATE">=TO_DATE(' 2013-10-27 00:00:00', 'syyyy-mm-dd hh24:mi:ss')) OR LNNVL("J_OP_TEST_ANML"."START_DATE"<=TO_DATE(' 2013-11-05
 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))))
86 - access("J_OP_TEST_ANML"."TEST_ID"="B_FP_TEST"."TEST_ID")
87 - filter("B_FP_TEST"."DATASOURCE_ID"=9 AND "B_FP_TEST"."IS_DELETED"=0 AND "B_FP_TEST"."TEST_NUM"<>1 AND "B_FP_TEST"."TEST_NUM"<>2 AND
 "B_FP_TEST"."TEST_NUM"<>99)
89 - filter("J_FP_INVESTIGATOR"."IS_PI"=1)
93 - filter("L_TEST_STATUS"."STATUS"='Completed' OR "L_TEST_STATUS"."STATUS"='In-Progress' OR "L_TEST_STATUS"."STATUS"='Scheduled')
94 - access("B_FP_TEST"."TEST_TYPE_ID"="L_ATMS_TEST_TYPE"."ATMS_TEST_TYPE_ID")
95 - filter("L_ATMS_TEST_TYPE"."IS_DELETED"=0)

Excellent piece of follow-up on my first suggestion.
I nearly made a comment about how the plan doesn't show Bloom filter pruning either - and then I realised why not. The plan you've shown us comes from Explain Plan with literal values present; the trace file shows bind variables with names that are generated when cursor_sharing is set to force or similar - so the run-time plan and the plan from explain plan are almost guaranteed to be different.
Oracle support will need you to supply the plan you get from trying to run the query and then making a call to dbms_xplan.display_cursor() - dbms_xplan in 10g | Oracle Scratchpad If you do this I think you'll find that the pstart/pstop columns contain entries like :BF0000, and you may even find operations link PX JOIN FILTER CREATE / PX JOIN FILTER USE
A couple of generic notes:
if a query does sufficient work to merit parallel execution, then it's usually better to supply the best possible information to the optimizer, which means using literals rather than bind variables - you could try executing the query with the hint /*+ cursor_sharing_exact */ to stop Oracle from turning your literals into binds; it might be the presence of bind variables that's making the optimizer choose a path that has to include bloom filter pruning in your case.
Where you have the to_date() call you've used a four-digit year - which is a very good thing and helps the optimizer - but it's also a good idea to include an explicit format string as well: with a four-digit year this probably won't make any difference, but it avoids any risk of ambiguity for the optimizer.
I made a comment about the P->S stage and bottlenecks - I spent a couple more minutes looking at the plan, and I see the optimizer has used concatentation: in effect it has run three query blocks one after the other and fed the results to the query co-ordinator - in this case the P->S would make no difference to the end-user response time there's always a final P->S to the coordinator, you just happen to have three of them.
Regards
Jonathan Lewis

Parallel processing in background

Hi All,
I am processing 1 million of records in background, which takes approximately around 10 hrs. I wanted to reduce the time to less than 1 hr and tried using parallel processing. But the tasks run in Dialog workprocesses and giving abap short dumps due to time out.
Is there any other solutions using that i can reduce total processing time.
Please note that i cannot split. I am getting 1 million records from a select query and after processing all those records in SAP, I am sending to XI and XI will post in legacy system.
Please note that all other performance tunings done.
Thanks,
Rajesh.

Hi Rajesh,
Refer sample code for Parallel Processing:
By doing this your processing time will be highly optimized.
Go thru the description given in the code at each level.
This code Checks available WORK PROCESSes and assigns data in packets for processing. This way you save a lot of time esp when data is in Millions.
Hope it helps.
REPORT PARAJOB.
Data declarations
DATA: GROUP LIKE RZLLITAB-CLASSNAME VALUE ' ',
"Parallel processing group.
"SPACE = group default (all
"servers)
WP_AVAILABLE TYPE I, "Number of dialog work processes
"available for parallel processing
"(free work processes)
WP_TOTAL TYPE I, "Total number of dialog work
"processes in the group
MSG(80) VALUE SPACE, "Container for error message in
"case of remote RFC exception.
INFO LIKE RFCSI, C, "Message text
JOBS TYPE I VALUE 10, "Number of parallel jobs
SND_JOBS TYPE I VALUE 1, "Work packets sent for processing
RCV_JOBS TYPE I VALUE 1, "Work packet replies received
EXCP_FLAG(1) TYPE C, "Number of RESOURCE_FAILUREs
TASKNAME(4) TYPE N VALUE '0001', "Task name (name of
"parallel processing work unit)
BEGIN OF TASKLIST OCCURS 10, "Task administration
TASKNAME(4) TYPE C,
RFCDEST LIKE RFCSI-RFCDEST,
RFCHOST LIKE RFCSI-RFCHOST,
END OF TASKLIST.
Optional call to SBPT_INITIALIZE to check the
group in which parallel processing is to take place.
Could be used to optimize sizing of work packets
work / WP_AVAILABLE).
CALL FUNCTION 'SPBT_INITIALIZE'
EXPORTING
GROUP_NAME = GROUP
"Name of group to check
IMPORTING
MAX_PBT_WPS = WP_TOTAL
"Total number of dialog work
"processes available in group
"for parallel processing
FREE_PBT_WPS = WP_AVAILABLE
"Number of work processes
"available in group for
"parallel processing at this
"moment
EXCEPTIONS
INVALID_GROUP_NAME = 1
"Incorrect group name; RFC
"group not defined. See
"transaction RZ12
INTERNAL_ERROR = 2
"R/3 System error; see the
"system log (transaction
"SM21) for diagnostic info
PBT_ENV_ALREADY_INITIALIZED = 3
"Function module may be
"called only once; is called
"automatically by R/3 if you
"do not call before starting
"parallel processing
CURRENTLY_NO_RESOURCES_AVAIL = 4
"No dialog work processes
"in the group are available;
"they are busy or server load
"is too high
NO_PBT_RESOURCES_FOUND = 5
"No servers in the group
"met the criteria of >
"two work processes
"defined.
CANT_INIT_DIFFERENT_PBT_GROUPS = 6
"You have already initialized
"one group and have now tried
"initialize a different group.
OTHERS = 7..
CASE SY-SUBRC.
WHEN 0.
"Everythings ok. Optionally set up for optimizing size of
"work packets.
WHEN 1.
"Non-existent group name. Stop report.
MESSAGE E836. "Group not defined.
WHEN 2.
"System error. Stop and check system log for error
"analysis.
WHEN 3.
"Programming error. Stop and correct program.
MESSAGE E833. "PBT environment was already initialized.
WHEN 4.
"No resources: this may be a temporary problem. You
"may wish to pause briefly and repeat the call. Otherwise
"check your RFC group administration: Group defined
"in accordance with your requirements?
MESSAGE E837. "All servers currently busy.
WHEN 5.
"Check your servers, network, operation modes.
WHEN 6.
Do parallel processing. Use CALL FUNCTION STARTING NEW TASK
DESTINATION IN GROUP to call the function module that does the
work. Make a call for each record that is to be processed, or
divide the records into work packets. In each case, provide the
set of records as an internal table in the CALL FUNCTION
keyword (EXPORT, TABLES arguments).
DO.
CALL FUNCTION 'RFC_SYSTEM_INFO' "Function module to perform
"in parallel
STARTING NEW TASK TASKNAME "Name for identifying this
"RFC call
DESTINATION IN GROUP group "Name of group of servers to
"use for parallel processing.
"Enter group name exactly
"as it appears in transaction
"RZ12 (all caps). You may
"use only one group name in a
"particular ABAP program.
PERFORMING RETURN_INFO ON END OF TASK
"This form is called when the
"RFC call completes. It can
"collect IMPORT and TABLES
"parameters from the called
"function with RECEIVE.
EXCEPTIONS
COMMUNICATION_FAILURE = 1 MESSAGE msg
"Destination server not
"reached or communication
"interrupted. MESSAGE msg
"captures any message
"returned with this
"exception (E or A messages
"from the called FM, for
"example. After exception
"1 or 2, instead of aborting
"your program, you could use
"SPBT_GET_PP_DESTINATION and
"SPBT_DO_NOT_USE_SERVER to
"exclude this server from
"further parallel processing.
"You could then re-try this
"call using a different
"server.
SYSTEM_FAILURE = 2 MESSAGE msg
"Program or other internal
"R/3 error. MESSAGE msg
"captures any message
"returned with this
"exception.
RESOURCE_FAILURE = 3. "No work processes are
"currently available. Your
"program MUST handle this
"exception.
YOUR_EXCEPTIONS = X. "Add exceptions generated by
"the called function module
"here. Exceptions are
"returned to you and you can
"respond to them here.
CASE SY-SUBRC.
WHEN 0.
"Administration of asynchronous RFC tasks
"Save name of task...
TASKLIST-TASKNAME = TASKNAME.
"... and get server that is performing RFC call.
CALL FUNCTION 'SPBT_GET_PP_DESTINATION'
EXPORTING
RFCDEST = TASKLIST-RFCDEST
EXCEPTIONS
OTHERS = 1.
APPEND TASKLIST.
WRITE: / 'Started task: ', TASKLIST-TASKNAME COLOR 2.
TASKNAME = TASKNAME + 1.
SND_JOBS = SND_JOBS + 1.
"Mechanism for determining when to leave the loop. Here, a
"simple counter of the number of parallel processing tasks.
"In production use, you would end the loop when you have
"finished dispatching the data that is to be processed.
JOBS = JOBS - 1. "Number of existing jobs
IF JOBS = 0.
EXIT. "Job processing finished
ENDIF.
WHEN 1 OR 2.
"Handle communication and system failure. Your program must
"catch these exceptions and arrange for a recoverable
"termination of the background processing job.
"Recommendation: Log the data that has been processed when
"an RFC task is started and when it returns, so that the
"job can be restarted with unprocessed data.
WRITE msg.
"Remove server from further consideration for
"parallel processing tasks in this program.
"Get name of server just called...
CALL FUNCTION 'SPBT_GET_PP_DESTINATION'
EXPORTING
RFCDEST = TASKLIST-RFCDEST
EXCEPTIONS
OTHERS = 1.
"Then remove from list of available servers.
CALL FUNCTION 'SPBT_DO_NOT_USE_SERVER'
IMPORTING
SERVERNAME = TASKLIST-RFCDEST
EXCEPTIONS
INVALID_SERVER_NAME = 1
NO_MORE_RESOURCES_LEFT = 2
"No servers left in group.
PBT_ENV_NOT_INITIALIZED_YET = 3
OTHERS = 4.
WHEN 3.
"No resources (dialog work processes) available at
"present. You need to handle this exception, waiting
"and repeating the CALL FUNCTION until processing
"can continue or it is apparent that there is a
"problem that prevents continuation.
MESSAGE I837. "All servers currently busy.
"Wait for replies to asynchronous RFC calls. Each
"reply should make a dialog work process available again.
IF EXCP_FLAG = SPACE.
EXCP_FLAG = 'X'.
"First attempt at RESOURCE_FAILURE handling. Wait
"until all RFC calls have returned or up to 1 second.
"Then repeat CALL FUNCTION.
WAIT UNTIL RCV_JOBS >= SND_JOBS UP TO '1' SECONDS.
ELSE.
"Second attempt at RESOURCE_FAILURE handling
WAIT UNTIL RCV_JOBS >= SND_JOBS UP TO '5' SECONDS.
"SY-SUBRC 0 from WAIT shows that replies have returned.
"The resource problem was therefore probably temporary
"and due to the workload. A non-zero RC suggests that
"no RFC calls have been completed, and there may be
"problems.
IF SY-SUBRC = 0.
CLEAR EXCP_FLAG.
ELSE. "No replies
"Endless loop handling
ENDIF.
ENDIF.
ENDCASE.
ENDDO.
Wait for end of job: replies from all RFC tasks.
Receive remaining asynchronous replies
WAIT UNTIL RCV_JOBS >= SND_JOBS.
LOOP AT TASKLIST.
WRITE:/ 'Received task:', TASKLIST-TASKNAME COLOR 1,
30 'Destination: ', TASKLIST-RFCDEST COLOR 1.
ENDLOOP.
This routine is triggered when an RFC call completes and
returns. The routine uses RECEIVE to collect IMPORT and TABLE
data from the RFC function module.
Note that the WRITE keyword is not supported in asynchronous
RFC. If you need to generate a list, then your RFC function
module should return the list data in an internal table. You
can then collect this data and output the list at the conclusion
of processing.
FORM RETURN_INFO USING TASKNAME.
DATA: INFO_RFCDEST LIKE TASKLIST-RFCDEST.
RECEIVE RESULTS FROM FUNCTION 'RFC_SYSTEM_INFO'
IMPORTING RFCSI_EXPORT = INFO
EXCEPTIONS
COMMUNICATION_FAILURE = 1
SYSTEM_FAILURE = 2.
RCV_JOBS = RCV_JOBS + 1. "Receiving data
IF SY-SUBRC NE 0.
Handle communication and system failure
ELSE.
READ TABLE TASKLIST WITH KEY TASKNAME = TASKNAME.
IF SY-SUBRC = 0. "Register data
TASKLIST-RFCHOST = INFO_RFCHOST.
MODIFY TASKLIST INDEX SY-TABIX.
ENDIF.
ENDIF.
ENDFORM
Reward points if that helps.
Manish
Message was edited by:
Manish Kumar

Parallel processing in large_pool

How can you set the parallel processes to use the large_pool instead of the shared_pool on an 11gR2 database? Pre 10g, parallel_automatic_tuning would set the large_pool for processing by paralel processes. Pparallel_automatic_tuning is deprecated since 10g. Also, I know that setting the PXuse_large_pool enables parallel processing in the large_pool. If PXuse_large_pool and parallel_automatic_tuning is set to FALSE and ASSM is disabled (sga_target=0), then what else can cause parallel processing to use large_pool?

Large Pool Sizing
Sizing the large pool can be complex. The large pool, if configured, must be at least 600 kilobytes in size. Usually, for most MTS applications, 600 is enough. However, if PQO (parallel query option) is also used in your Oracle8,Oracle8i, or Oracle9i environment, then the size of the large pool will increase dramatically. The V$SGASTAT dynamic performance view has a new column in Oracle8 called POOL. It is used to contain the pool area where that particular type of object is being stored. By issuing a summation select against the V$SGASTAT view, a DBA can quickly determine the size of the large pool area currently being used.
SELECT name, SUM(bytes) FROM V$SGASTAT WHERE pool='LARGE POOL' GROUP BY ROLLUP(name);
The above select should be used when an “ORA-04031:Unable to allocate 16084 bytes of shared memory (‘large pool’, ‘unknown object’, ‘large pool hea’, ‘PX large pool’)” error is received during operation with a large pool configured (the number of bytes specified may differ). When the above select is run, the resulting summary number of bytes will indicate the current size of the pool and show how close you are to your maximum, as specified in the initialization parameter LARGE_POOL_SIZE. Generally, increasing the large_pool by up to 100 percent will eliminate the ORA-04031 errors.
In an undocumented (as of this writing) bug in 8.1.7.1 if you have large complex SQL statements, such as a 15 table join, and nothing to force large pool usage you can get excessive CPU times, ORA-04031 and ORA_01037 errors. If you get ORA-04031, or ORA-01037 errors and aren't using the large pool try turning on parallel query by setting the parallel query processes to minimum values to force use of the large pool.

Performance in Parallel Processing

Hi,
My program model is as follows:
I have a RFC FM(ZRFC_BACKGROUND) that will be called in BACKGROUND TASK and
This FM in turn will call a child process(another RFC FM(ZRFC_CHILD)) in parallel for packages that are created in ZRFC_BACKGROUND
We can use around 10 work processes that belongs to one server group, for parallel processing
Facts regarding Packaging:
One Package has 4 retry attempts due to locking issues.if there is an error in child process, the package is held back in FM : ZRFC_BACKGROUND and re triggered after sometime
Locking issues are frequent.
Data in DB tables on which I will put SELECT query:
Header        Table LIKP (With approximately 1 Million Entries) (My select query will fetch around    800 entries)
Item             Table LIPS (With approximately 6 Million Entries) (My select query will fetch around 10000 entries)
Partner        Table VBPA (With around 3 Million Entries) (My select query will fetch around 3000 entries)
Reference   Table LIPSRF (With around 74 million entries) (My select query will fetch around 10000 entries)
My QUESTION is which of the following approach is better:
     To put LIKP select query in the first background FM ZRFC_BACKGROUND and pass it to ZRFC_CHILD and do remaining query in ZRFC_CHILD
     OR to put all the select query in ZRFC_BACKGROUND
Please keep following things in mind :
As we know it is not recommended to pass huge tables in RFC as it may increase the network load (In case of select query in ZRFC_BACKGROUND, network load will be more)
But it is also not recommended to do DB select several times (In case of select query in ZRFC_CHILD, db read time will be more )
On the basis of aove facts, please suggest which is the better approach.
Regards
Rocky

Hi Rocky,
Why do you have locking issues? In the worst case, parallel processing will not help at all or even slow down the process if there are too many locks. Ideally, parallel tasks should not collide with each other. Which tables do you update / insert?
How big are your packages? How long do they take to process?
From the database point of view, it will not make a big difference whether you select 1 x 10000 entries or 100 x 100 entries. On the other side, if your second approach leads to additional database accesses as compared to the first one, this of course could slow down the process.
Best Regards, Randolf

Basic query regarding work-area and select query

hi
dear sdn members,
thanks too all for solving all my query's up till now
i am stuck in a problem need help
1) why basically work-area has been used ? the sole purpose
2) different types of select query ? only coding examples
note: no links pls
regards,
virus

hi,
Work Area
Description for a data object that is particularly useful when working with internal tables or database tables as a source for changing operations or a target for reading operations.
WORKAREA is a structure that can hold only one record at a time. It is a collection of fields. We use workarea as we cannot directly read from a table. In order to interact with a table we need workarea. When a Select Statement is executed on a table then the first record is read and put into the header of the table and from there put into the header or the workarea(of the same structure as that of the table)of the internal table and then transferred top the body of the internal table or directly displayed from the workarea.
Each row in a table is a record and each column is a field.
While adding or retrieving records to / from internal table we have to keep the record temporarily.
The area where this record is kept is called as work area for the internal table. The area must have the same structure as that of internal table. An internal table consists of a body and an optional header line.
Header line is a implicit work area for the internal table. It depends on how the internal table is declared that the itab will have the header line or not.
.g.
data: begin of itab occurs 10,
ab type c,
cd type i,
end of itab. " this table will have the header line.
data: wa_itab like itab. " explicit work area for itab
data: itab1 like itab occurs 10. " table is without header line.
The header line is a field string with the same structure as a row of the body, but it can only hold a single row.
It is a buffer used to hold each record before it is added or each record as it is retrieved from the internal table. It is the default work area for the internal table.
With header line
SELECT.
Put the curson on that word and press F1 . You can see the whole documentation for select statements.
select statements :
SELECT result
FROM source
INTO|APPENDING target
[[FOR ALL ENTRIES IN itab] WHERE sql_cond]
Effect
SELECT is an Open-SQL-statement for reading data from one or several database tables into data objects.
The select statement reads a result set (whose structure is determined in result ) from the database tables specified in source, and assigns the data from the result set to the data objects specified in target. You can restrict the result set using the WHERE addition. The addition GROUP BY compresses several database rows into a single row of the result set. The addition HAVING restricts the compressed rows. The addition ORDER BY sorts the result set.
The data objects specified in target must match the result set result. This means that the result set is either assigned to the data objects in one step, or by row, or by packets of rows. In the second and third case, the SELECT statement opens a loop, which which must be closed using ENDSELECT. For every loop pass, the SELECT-statement assigns a row or a packet of rows to the data objects specified in target. If the last row was assigned or if the result set is empty, then SELECT branches to ENDSELECT . A database cursor is opened implicitly to process a SELECT-loop, and is closed again when the loop is ended. You can end the loop using the statements from section leave loops.
Up to the INTO resp. APPENDING addition, the entries in the SELECTstatement define which data should be read by the database in which form. This requirement is translated in the database interface for the database system´s programming interface and is then passed to the database system. The data are read in packets by the database and are transported to the application server by the database server. On the application server, the data are transferred to the ABAP program´s data objects in accordance with the data specified in the INTO and APPENDING additions.
System Fields
The SELECT statement sets the values of the system fields sy-subrc and sy-dbcnt.
sy-subrc Relevance
0 The SELECT statement sets sy-subrc to 0 for every pass by value to an ABAP data object. The ENDSELECT statement sets sy-subrc to 0 if at least one row was transferred in the SELECT loop.
4 The SELECT statement sets sy-subrc to 4 if the result set is empty, that is, if no data was found in the database.
8 The SELECT statement sets sy-subrc to 8 if the FOR UPDATE addition is used in result, without the primary key being specified fully after WHERE.
After every value that is transferred to an ABAP data object, the SELECT statement sets sy-dbcnt to the number of rows that were transferred. If the result set is empty, sy-dbcnt is set to 0.
Notes
Outside classes, you do not need to specify the target area with INTO or APPENDING if a single database table or a single view is specified statically after FROM, and a table work area dbtab was declared with the TABLES statement for the corresponding database table or view. In this case, the system supplements the SELECT-statement implicitly with the addition INTO dbtab.
Although the WHERE-condition is optional, you should always specify it for performance reasons, and the result set should not be restricted on the application server.
SELECT-loops can be nested. For performance reasons, you should check whether a join or a sub-query would be more effective.
Within a SELECT-loop you cannot execute any statements that lead to a database commit and consequently cause the corresponding database cursor to close.
SELECT - result
Syntax
... lines columns ... .
Effect
The data in result defines whether the resulting set consists of multiple rows (table-like structure) or a single row ( flat structure). It specifies the columns to be read and defines their names in the resulting set. Note that column names from the database table can be changed. For single columns, aggregate expressions can be used to specify aggregates. Identical rows in the resulting set can be excluded, and individual rows can be protected from parallel changes by another program.
The data in result consists of data for the rows lines and for the columns columns.
SELECT - lines
Syntax
... { SINGLE }
| { { } } ... .
Alternatives:
1. ... SINGLE
2. ... { }
Effect
The data in lines specifies that the resulting set has either multiple lines or a single line.
Alternative 1
... SINGLE
Effect
If SINGLE is specified, the resulting set has a single line. If the remaining additions to the SELECT command select more than one line from the database, the first line that is found is entered into the resulting set. The data objects specified after INTO may not be internal tables, and the APPENDING addition may not be used.
An exclusive lock can be set for this line using the FOR UPDATE addition when a single line is being read with SINGLE. The SELECT command is used in this case only if all primary key fields in logical expressions linked by AND are checked to make sure they are the same in the WHERE condition. Otherwise, the resulting set is empty and sy-subrc is set to 8. If the lock causes a deadlock, an exception occurs. If the FOR UPDATE addition is used, the SELECT command circumvents SAP buffering.
Note
When SINGLE is being specified, the lines to be read should be clearly specified in the WHERE condition, for the sake of efficiency. When the data is read from a database table, the system does this by specifying comparison values for the primary key.
Alternative 2
Effect
If SINGLE is not specified and if columns does not contain only aggregate expressions, the resulting set has multiple lines. All database lines that are selected by the remaining additions of the SELECT command are included in the resulting list. If the ORDER BY addition is not used, the order of the lines in the resulting list is not defined and, if the same SELECT command is executed multiple times, the order may be different each time. A data object specified after INTO can be an internal table and the APPENDING addition can be used. If no internal table is specified after INTO or APPENDING, the SELECT command triggers a loop that has to be closed using ENDSELECT.
If multiple lines are read without SINGLE, the DISTINCT addition can be used to exclude duplicate lines from the resulting list. If DISTINCT is used, the SELECT command circumvents SAP buffering. DISTINCT cannot be used in the following situations:
If a column specified in columns has the type STRING, RAWSTRING, LCHAR or LRAW
If the system tries to access pool or cluster tables and single columns are specified in columns.
Note
When specifying DISTINCT, note that you have to carry out sort operations in the database system for this.
SELECT - columns
Syntax
| { {col1|aggregate( col1 )}
{col2|aggregate( col2 )} ... }
| (column_syntax) ... .
Alternatives:
1. ... *
2. ... {col1|aggregate( col1 )}
{col2|aggregate( col2 )} ...
3. ... (column_syntax)
Effect
The input in columns determines which columns are used to build the resulting set.
Alternative 1
Effect
If * is specified, the resulting set is built based on all columns in the database tables or views specified after FROM, in the order given there. The columns in the resulting set take on the name and data type from the database tables or views. Only one data object can be specified after INTO.
Note
If multiple database tables are specified after FROM, you cannot prevent multiple columns from getting the same name when you specify *.
Alternative 2
... {col1|aggregate( col1 )}
{col2|aggregate( col2 )} ...
Effect
A list of column labels col1 col2 ... is specified in order to build the resulting list from individual columns. An individual column can be specified directly or as an argument of an aggregate function aggregate. The order in which the column labels are specified is up to you and defines the order of the columns in the resulting list. Only if a column of the type LCHAR or LRAW is listed does the corresponding length field also have to be specified directly before it. An individual column can be specified multiple times.
The addition AS can be used to define an alternative column name a1 a2 ... with a maximum of fourteen digits in the resulting set for every column label col1 col2 .... The system uses the alternative column name in the additions INTO|APPENDING CORRESPONDING FIELDS and ORDER BY. .
Column labels
The following column labels are possible:
If only a single database table or a single view is specified after FROM, the column labels in the database table - that is, the names of the components comp1 comp2... - can be specified directly for col1 col2 ... in the structure of the ABAP Dictionary.
If the name of the component occurs in multiple database tables of the FROM addition, but the desired database table or the view dbtab is only specified once after FROM, the names dbtab~comp1 dbtab~comp2 ... have to be specified for col1 col2 .... comp1 comp2 ... are the names of the components in the structure of the ABAP Dictionary.
If the desired database table or view occurs multiple times after FROM, the names tabalias~comp1 tabalias~comp2 ... have to be specified for col1 col2 .... tabalias is the alternative table name of the database table or view defined after FROM, and comp1 comp2 ... are the names of the components in the structure of the ABAP Dictionary.
The data type of a single column in the resulting list is the datatype of the corresponding component in the ABAP Dictionary. The corresponding data object after INTO or APPENDING has to be selected accordingly.
Note
If multiple database tables are specified after FROM, you can use alternative names when specifying single columns to avoid having multiple columns with the same name.
Example
Read specific columns of a single row.
DATA wa TYPE spfli.
SELECT SINGLE carrid connid cityfrom cityto
INTO CORRESPONDING FIELDS OF wa
FROM spfli
WHERE carrid EQ 'LH' AND connid EQ '0400'.
IF sy-subrc EQ 0.
WRITE: / wa-carrid, wa-connid, wa-cityfrom, wa-cityto.
ENDIF.
Alternative 3
... (column_syntax)
Effect
Instead of static data, a data object column_syntax in brackets can be specified, which, when the command is executed, either contains the syntax shown with the static data, or is initial. The data object column_syntax can be a character-type data object or an internal table with a character-type data type. The syntax in column_syntax, like in the ABAP editor, is not case-sensitive. When specifying an internal table, you can distribute the syntax over multiple rows.
If column_syntax is initial when the command is executed, columns is implicitly set to * and all columns are read.
If columns are specificied dynamically without the SINGLE addition, the resulting set is always regarded as having multiple rows.
Notes
Before Release 6.10, you could only specify an internal table with a flat character-type row type for column_syntax with a maximum of 72 characters. Also, before Release 6.10, if you used the DISTINCT addition for dynamic access to pool tables or cluster tables, this was ignored, but since release 6.10, this causes a known exception.
If column_syntax is an internal table with header line, the table body and not the header line is evaluated.
Example
Read out how many flights go to and from a city. The SELECT command is implemented only once in a sub-program. The column data, including aggregate function and the data after GROUP BY, is dynamic. Instead of adding the column data to an internal l_columns table, you could just as easily concatenate it in a character-type l_columns field.
PERFORM my_select USING `CITYFROM`.
ULINE.
PERFORM my_select USING `CITYTO`.
FORM my_select USING l_group TYPE string.
DATA: l_columns TYPE TABLE OF string,
l_container TYPE string,
l_count TYPE i.
APPEND l_group TO l_columns.
APPEND `count( * )` TO l_columns.
SELECT (l_columns)
FROM spfli
INTO (l_container, l_count)
GROUP BY (l_group).
WRITE: / l_count, l_container.
ENDSELECT.
ENDFORM.
SELECT - aggregate
Syntax
... { MAX( col )
| MIN( col )
| AVG( col )
| SUM( col )
| COUNT( DISTINCT col )
| COUNT( * )
| count(*) } ... .
Effect
As many of the specified column labels as you like can be listed in the SELECT command as arguments of the above aggregate expression. In aggregate expressions, a single value is calculated from the values of multiple rows in a column as follows (note that the addition DISTINCT excludes double values from the calculation):
MAX( col ) Determines the maximum value of the value in the column col in the resulting set or in the current group.
MIN( col ) Determines the minimum value of the content of the column col in the resulting set or in the current group.
AVG( col ) Determines the average value of the content of the column col in the resulting set or in the current group. The data type of the column has to be numerical.
SUM( col ) Determines the sum of the content of the column col in the resulting set or in the current group. The data type of the column has to be numerical.
COUNT( DISTINCT col ) Determines the number of different values in the column col in the resulting set or in the current group.
COUNT( * ) (or count(*)) Determines the number of rows in the resulting set or in the current group. No column label is specified in this case.
If you are using aggregate expressions, all column labels that are not listed as an argument of an aggregate function are listed after the addition GROUP BY. The aggregate functions evaluate the content of the groups defined by GROUP BY in the database system and transfer the result to the combined rows of the resulting set.
The data type of aggregate expressions with the function MAX, MIN or SUM is the data type of the corresponding column in the ABAP Dictionary. Aggregate expressions with the function AVG have the data type FLTP, and those with COUNT have the data type INT4. The corresponding data object after INTO or APPENDING has to be selected accordingly.
Note the following points when using aggregate expressions:
If the addition FOR ALL ENTRIES is used in front of WHERE, or if cluster or pool tables are listed after FROM, no other aggregate expressions apart from COUNT( * ) can be used.
Columns of the type STRING or RAWSTRING cannot be used with aggregate functions.
When aggregate expressions are used, the SELECT command makes it unnecessary to use SAP buffering.
Null values are not included in the calculation for the aggregate functions. The result is a null value only if all the rows in the column in question contain the null value.
If only aggregate expressions are used after SELECT, the results set has one row and the addition GROUP BY is not necessary. If a non-table type target area is specified after INTO, the command ENDSELECT cannot be used together with the addition SINGLE. If the aggregate expression count( * ) is not being used, an internal table can be specified after INTO, and the first row of this table is filled.
If aggregate functions are used without GROUP BY being specified at the same time, the resulting set also contains a row if no data is found in the database. If count( * ) is used, the column in question contains the value 0. The columns in the other aggregate functions contain initial values. This row is assigned to the data object specified after INTO, and unless count( * ) is being used exclusively, sy-subrc is set to 0 and sy-dbcnt is set to 1. If count( *) is used exclusively, the addition INTO can be omitted and if no data can be found in the database, sy-subrc is set to 4 and sy-dbcnt is set to 0.
if helpful reward points

Parallel processing of Jobs in Java

Hi Sun forum Experts,
we are facing a problem in parallel processing of jobs in java,The jdk version used is 1.4.
Find the issue analysis as below
Assume we have TABLE1 as below:
1st job completed running and updates the status to OK by running an update query (so STATUS in TABLE1 is OK )---> 1st row
2nd job is still running and is in IN (Status in Table) ---> 2nd row
3rd job is still running and is in IN (Status in Table) ---> 3rd row
4th job is still running and is in IN (Status in Table) ---> 4th row
In the meanwhile, the First job has completed the job and hence status is "OK" now. now we are calling another update query which updates another column in another table(TABLE2) in database.
The query is " Update TABLE2 set STATUS="OK" where INIT_NO='123456789' AND NOT EXISTS(SELECT COUNT(*) FROM TABLE1 WHERE STATUS in('IN','RE'))
Hence the subquery is returning 3 rows as count for the inner query execution which skips the updation of the TABLE2.
{code}
*The problem is we want all the 4 jobs to be completed first with respect to TABLE1(running a update query)   & Once this is done we need to update another column in another table(TABLE2) in database.*
i.e as shown below
{code}
public void execute(){
             dao.updateTable1(); // calls Table1 and updates it to OK. --> This has to be performed for n number for jobs.Finish all job completion and then proceed below.
            //Once UpdateTable1 is completed ,after this we perform the below step
            dao.updateTable2(); //calls Table2 and updates it to OK
{code}
how do we handle this scenario in Java,
Please Note:Version in java is used is 1.4 only
Deepak

yes sir,you are absolutely correct.
We run 4 jobs to update table1, wait for them all to complete, then update table2   can you please tell me how do we achieve this in java,
Please note:JDK version is 1.4 which we are using.
Thanks
Deepak
Edited by: Deepak_A_L on Jun 25, 2009 12:58 AM

How to monitor parallel processing

Hi All,
I have 2 short questions:
1. When we want parallel processing, we can either use a parallel hing in the query, or alter a table to be parrallel. My question is what is the difference in the following 2 syntax:
a. ALTER TABLE myTable PARALLEL (DEGREE 3);
b. ALTER TABLE myTable PARALLEL 3;
Does the "DEGREE" keywor make any difference? or they both are same statements?
2. When we enable parallel processing, how can we monitor oracle processes, to confirm that a certain table is being actually processed by multiple threads of a singe user process?
An early response would be highly appreciated. Thanks.

user566817 wrote:
2. When we enable parallel processing, how can we monitor oracle processes, to confirm that a certain table is being actually processed by multiple threads of a singe user process?There are a number of virtual performance views that can be used. Please refer to the Oracle® Database Reference guide for details on these.
Had a look though my scripts and I have this one.. cannot recall if I "borrowed" it from somewhere and customised it and how old it is.. but it should (hopefully) still be mostly correct. It uses the virtual view v$px_process to determine the list of current PQ slaves in the pool and if they are used, map them to the Oracle session using them.
select distinct
        x.server_name           as "PQ",
        x.status                as "Status",
        x.sid                   as "OraPID",
        w2.sid                  as "Parent OraPID",
        v.osuser                as "O/S User",
        v.schemaname            as "User",
        w1.event                as "Child Wait",
        w2.event                as "Parent Wait"
from    v$px_process    x,
        v$lock          l,
        v$session       v,
        v$session_wait w1,
        v$session_wait w2
where   x.sid =! l.sid(+)
and     to_number (substr(x.server_name,3)) = l.id2(+)
and     x.sid = w1.sid(+)
and     l.sid = w2.sid(+)
and     x.sid = v.sid(+)
and     nvl(l.type,'PS') = 'PS'Use at own risk - best would be to verify that this is still valid using the Reference Guide. Or create similar ones using the available V$ views for the details you want to see (e.g. SQL statement executed per PQ, etc).

Parallel process in select query

Similar Messages

Maybe you are looking for