the update statement modidy
Hi all
I have here is the update statement in my code
UPDATE equip_stg stg
SET stg.err = 'WRC '.
WHERE DOES NOT EXIST (SELECT *)
FROM (SELECT stg.*
Of equip_stg stg.
REASON_SETUP sea
WHERE stg.equipment_fk = mer.equipment_pk
AND stg.downtime_reason_code = mer.reason_code
AND ((stg.status = 3 AND sea. REASON_TYPE = 1)
OR (stg.status = 2 AND sea. REASON_TYPE = 3)
OR (stg.status = 1 AND sea. REASON_TYPE = 4))
ERS stg.downtime_reason_code AND IS NOT NULL)
WHERE ers.equipment_fk = stg.equipment_fk
AND ers.downtime_reason_code = stg.downtime_reason_code
)
AND stg.status IN (3,2,1)
AND stg.downtime_reason_code IS NOT NULL;
The statement above sets updated the column equip_stg in WRC if err for corresponding material and reason, there is no line in the REASON_SETUP table.
I want to change the code above so that the validation above is first done with REASON_SETUP according to equipment and column of type of reason. Is there is no line for this type of material and the reason in this table, then he must hit lookup table. Example, I have data from equipment EQP1, status = 3, then it should check if REASON_SETUP has everything save for EQP1 equipment and the reasoning of type 1. If she made the next record should be validated on the table otherwise come file should be validated against the lookup table. For the table of choice also I have same join conditions
STG.equipment_fk = lookup.equipment_pk
AND stg.downtime_reason_code = lookup.reason_code
Thanks in advance.
Kind regards
Amrit
Hello
I think I see. All lines for each (equipment, reason_type) will be removed completely on a table or another. reason_setup if possible and looking as if this combination never occurs in reason_type. Another way to say it is that you want only (equipment, reason_type) groups of the best tablle, where reason_type is better than search. Exactly what made a Request Top - N . Here's a way to do it:
UPDATE equip_stg
SET err_code = "WRC".
WHERE (equipment, downtime_reason_code) NOT IN
(
WITH union_data AS
(
SELECT 'A' preference AS, reason_code, reason_type, equipment
OF reason_setup
UNION ALL
SELECT equipment, reason_type, reason_code, 'B' AS preferences
SEARCH
)
got_r_num AS
(
SELECT equipment, reason_type, reason_code
DENSE_RANK () (PARTITION BY MATERIAL, reason_type
ORDER OF preference
) AS r_num
Of union_data
)
SELECT equipment
DECODE (reason_type
1, 3
3, 2
)
reason_code
OF got_r_num
WHERE r_num = 1
)
;
To see how this works, reconstruction of the NOT IN subquery, step by step. First run just union_data as a stand-alone application. When you understand what it does, run got_r_num and when you see what he's doing, run the set NOT IN subquery.
Tags: Database
Similar Questions
-
Commit to taking a lot of time for the update statement to bulk - 11 GR 2
Hi team,
We have a source table, which is the main table for our application.
When I commit a transaction after update block, it takes long about 2 to 4 minutes.
I don't understand why it takes too long...
Could you please help me on this fix, please...
It's the details on the table,
Total number of records: 35 M record
Validation interval: 500 records / commit
Total scores: 3
total number of columns: 95 - including primary, unique and foreign key columns (all columns will be updated for each update because update from the app online)
Total no of the foreign key columns: 12
Unique key column: 1
Total index: 27 ( including the key index foreign 12 + 1 + 1 index of primary key Unique key index )
example of update statement,
UPDATE DIRECTORY_LISTING_STG SET COLUMN_1 = :VAL_1 . . . COLUMN_94 = :VAL_94 WHERE RECORD_KEY = :P_KEY_VAL;
The table is.
Plan hash value: 2997337465 -------------------------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | -------------------------------------------------------------------------------------------- | 0 | UPDATE STATEMENT | | 1 | 308 | 3 (0)| 00:00:01 | | 1 | UPDATE | DIRECTORY_LISTING_STG | | | | | |* 2 | INDEX UNIQUE SCAN| XPKDLS_STG_PARTPN | 1 | 308 | 2 (0)| 00:00:01 | -------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 2 - access("RECORD_KEY"=TO_NUMBER(:P_KEY))Details of the database,
Oracle Database 11g Enterprise Edition Release 11.1.0.7.0 - 64bit Production PL/SQL Release 11.1.0.7.0 - Production "CORE 11.1.0.7.0 Production" TNS for Linux: Version 11.1.0.7.0 - Production NLSRTL Version 11.1.0.7.0 - Production
Thank you
RAM
2793617 wrote:
Hello
I am updating table via an application using Hibernate jdbc connection online.
If the update statement will be sent to the database in a pack of 500 documents (batch update).
Thanks Ram
line-by-line UPDATE is slow by slow transformation
In Oracle no other session can 'see' upgrade from uncommitted changes.
The only way to significantly reduce the elapsed time is to have only 1 UPDATE & 1 COMMIT!
-
line blocking during the update statement
HelloUsing Oracle 11 g 2. If I call the following update statement of meeting A, meeting A locks the line until a statement commit / rollback is emanating from the session has. If the session B calls the same statement was updated and the same line, session B will have to wait until the lock is released in the session. However, application developers are expressed in terms of discussions. Would it be possible that the update statement is called in the same session by several requests? If so, the case statement can be evaluated without the line being locked which could lead to erroneous results? Trying to keep this short message.
tableA has column (number of key, primary) and columnB (number)
{
Update tableA
Define columnB = case when columnB = 3 then
4
When columnB = 4 then
5
on the other
columnB
end
When columnA = 6;
}
2 applications (almost at the same time) in the same session could assess columnB 3. The goal would be the first query sets in column 4 and the second request sets in column 5.
> However, application developers are expressed in terms of discussions.
Of course, they are...
But each application thread will have its own dedicated database session.
-
How to use the user defined function in the Update statement
Hi all
I wrote under the update statement to update the column based on the return value of function. but it does not work. Could someone help me on this. This function returns a single value for each project.
Thanks in advance.
UPDATE dg2. OD_PROJ_LOOKUP_TEMP o
SET Months_In_Stage_Cnt = Months_In_Stage_Cnt_ret (o.project_id) select the double;
Thank you
DebPLS-00231: function 'MONTHS_IN_STAGE_CNT_RET' cannot be used in SQL
-What is the function that is declared private in a package? If Yes, then it must be public, otherwise the SQL engine cannot be used.
Published by: Dom Brooks on November 18, 2011 09:43
-
Decode function in the Update statement
Hi all
I'm writing a query where I can update a pastdue_fees column in a table of book_trans based on a difference between the return_dte and due_dte columns.
I use Oracle SQL. That's what I have so far for my decoding function:
SQL > SELECT
2 DECODE (SIGN ((return_dte-due_dte) * 2),)
3 '-1 ', ' 0',
4 '1', '12', 'Null')
5 FROM book_trans;
DECO
----
Null value
12
Null value
0
If the logic is that, if the sign is - 1, the value in the return_dte column must be 0; If it is + 1, then it has 12 and everything else is Null.
So now, I have to conclude my function of decoding of the update statement to update the columns. However, I get the error messages.
The logic should be:
UPDATE book_trans SET PastDue_fees = decode (expression)
I gave him a few different tests with the following results:
SQL > UPDATE book_trans
2 SET pastdue_fees = SELECT
3 DECODE (SIGN ((return_dte-due_dte) * 2),)
4 '-1 ', ' 0',
5 '1', '12', 'Null')
6 FROM book_trans.
SET pastdue_fees = SELECT
*
ERROR on line 2:
ORA-00936: lack of expression
SQL > UPDATE book_trans
2 SET pastdue_fees =
3 DECODE (SIGN ((return_dte-due_dte) * 2),)
4 '-1 ', ' 0',
5 '1', '12', 'Null')
6 FROM book_trans.
OF book_trans
*
ERROR on line 6:
ORA-00933: SQL not correctly completed command.
Any help or advice would be greatly appreciated I got SQL for about six weeks and not very competent!
Thank you!882300 wrote:
Hi allI'm writing a query where I can update a pastdue_fees column in a table of book_trans based on a difference between the return_dte and due_dte columns.
I use Oracle SQL. That's what I have so far for my decoding function:
SQL > SELECT
2 DECODE (SIGN ((return_dte-due_dte) * 2),)
3 '-1 ', ' 0',
4 '1', '12', 'Null')
5 FROM book_trans;DECO
----
Null value
12
Null value
0If the logic is that, if the sign is - 1, the value in the return_dte column must be 0; If it is + 1, then it has 12 and everything else is Null.
So now, I have to conclude my function of decoding of the update statement to update the columns. However, I get the error messages.
The logic should be:
UPDATE book_trans SET PastDue_fees = decode (expression)I gave him a few different tests with the following results:
SQL > UPDATE book_trans
2 SET pastdue_fees = SELECT
3 DECODE (SIGN ((return_dte-due_dte) * 2),)
4 '-1 ', ' 0',
5 '1', '12', 'Null')
6 FROM book_trans.
SET pastdue_fees = SELECT
*
ERROR on line 2:
ORA-00936: lack of expressionSQL > UPDATE book_trans
2 SET pastdue_fees =
3 DECODE (SIGN ((return_dte-due_dte) * 2),)
4 '-1 ', ' 0',
5 '1', '12', 'Null')
6 FROM book_trans.
OF book_trans
*
ERROR on line 6:
ORA-00933: SQL not correctly completed command.Any help or advice would be greatly appreciated I got SQL for about six weeks and not very competent!
Thank you!
If you really really really want to update the entire table, the syntax would be...
UPDATE book_trans SET pastdue_fees = DECODE(SIGN((return_dte - due_dte)*2), -1, 0, 1, 12, Null);I took out all the single quotes. If you actually have a string column and store all numbers in there so it must be reported as a NUMBER column and not a column of type character (varchar2).
ALWAYS use the appropriate data type, it will save you a ton of headaches in the future.
Also, since you are new to the forum, please read the FAQ to learn etiquette and what not.
http://wikis.Sun.com/display/Forums/Forums+FAQ
-
need help with the Update statement
Hello
I received a question in a course and I tried my best to respond, and now my brain is giving. I would really appreciate help with the update statement. I don't mind if you do not validate a solution, a little nudge in the right direction would be really useful. I'll post that I got.
THE QUESTION
/ * For these agents disabled on more than seven missions, change their date of deactivation of the first date of deactivation of all the agents that have been activated in the same year as the agent that you update currently.
*/
I have it divided into parts, here is my select statement to agents disabled on more than 7 missions, which produces the deactivation_dates in the agents table that I want to update...
.. .and the code for the first date of deactivation for each year of activation agentSELECT s.deactivation_date FROM ( SELECT a.deactivation_date, count(m.mission_id) as nomissions FROM agents a INNER JOIN missions_agents m on a.agent_id=m.agent_id GROUP BY a.deactivation_date ) s WHERE s.nomissions>7 AND s.deactivation_date IS NOT NULL
..... I am not real to marry these two statements together in the Update statement. I can't extract each date of deactivation produced in the first select statement and their match against the first date of deactivation in the year they have been activated for the second select statement.select a2.deactivation_date from agents a2 where a2.deactivation_date= ( select min(a.deactivation_date) from agents a where to_number(to_char(a.activation_date,'YYYY'))=to_number(to_char(a2.activation_date,'YYYY')) )
Any help greatly appreciated... :))I began to wonder how things would :)
user8695469 wrote:
First of all why he chooses the date the earliest of all agentsUPDATE AGENTS_COPY AC /* (1) */ SET DEACTIVATION_DATE = ( SELECT MIN(AGS.DEACTIVATION_DATE) FROM AGENTS_COPY AGS , AGENTS_COPY AC /* (2) */ WHERE TRUNC(AGS.ACTIVATION_DATE,'YEAR') = TRUNC(AC.ACTIVATION_DATE,'YEAR') /* (3) */ )He recovers as soon as the subquery has not been correctly set in the SET clause. It seems you are trying to update a correlated, but we are still having a conceptual shift. I have added a few comments to your code above and below will explain.
(1): when you do a correlated update it is useful to the table alias that you did right here.
(2): this table statement is not necessary and is the reason why the FIRST deactivation date is selected. The alias that you use (3) refers to THIS table, not the one defined in the update statement. Remove the line indicated by (2) in the FROM clause and a correlated update will happen.
and secondly why is it to update each row, when I thought that I'm just the lines where the agents are disabled and missions > 7? Pointers on where I'm wrong would be very appreciated. (SQL = stupid query language!) :)
user8695469 wrote: then why is it to update each row, when I thought that I'm just the lines where the agents are disabled and missions > 7? Pointers on where I'm wrong would be very appreciated. (SQL = stupid query language!) :)
WHERE EXISTS ( SELECT a.agent_id, count(m.mission_id) FROM agents a /* INNER JOIN AC ON AC.AGENT_ID = A.AGENT_ID */ INNER JOIN missions_agents m ON a.agent_id=m.agent_id GROUP BY a.agent_id, a.deactivation_date HAVING count(m.mission_id)>7 AND a.deactivation_date IS NOT NULL )Once again this problem is similar to the question above that a correlation update doesn't work. Test existence of lines in an EXISTS subquery. Since your subquery is not related to the table that you are trying to update, it will be always return a line and, therefore, it returns true for EACH LINE in the AGENTS table. To limit the game to only agents > 7 missions results, you need to add a join condition that references the table in your update statement. I added one above (with comments) as a sample.
I recommend you look over all material that you have associated with correlated subqueries, including documents that I posted above. This seems to be what you're having the problem more with. If you need me to explain the concept of correlated queries any better please let me know.
Thank you!
-
Helps when a special character and is included in the update statement
Hi gurus,
Please suggest me here. Your help is greatly appreciated.Am wririting an update statement on a column called Phone_num.description varchar2 (40).
Here is my update statement:
UPDATE Phone_num
DESCRIPTION SET = "Beta - EMV / not for the purpose of Production."
WHERE DIAL_NUMBER = 18662307528
AND DESCRIPTION =' to & VERIZON';
Here is the structure of the table:Phone_num. Description varchar2 (40)
Phone_num. DIAL_NUMBER Number (11)Sample data:
So when I run the sql script it pop-up is to highlight.Please suggest how I can handle this.
Hello
As John said, the easiest thing is just to disable the functionality of substitution of variables, saying:
SET DEFINE OFF
before the statement. If you need to re - enable this feature, then tell
ALL SET ON
just after the statement.
If you can't do that, then break the literal string in 2 parts immediately after the & character, like this:
UPDATE Phone_num
DESCRIPTION SET = "Beta - EMV / not for the purpose of Production."
WHERE DIAL_NUMBER = 18662307528
AND DESCRIPTION = ' to &'. | "VERIZON";
You can't have a substitution variable whose name starts by ', then SQL ("more does not interpret &" as meaning a substituton variable.)
-
How to optimize the update statement so that the query is reading the same table?
Hi all
I have a process with the following UPDATE statement:
Sales_val update
SET ship_date =)
Select max (hist.ship_date)FROM sales_hist hist
WHERE hist. X_ID = A.X_ID
AND hist. X_VAL = A.X_VAL)
) WHERE EXISTS (select * from sales_hist hist where )
WHERE hist. X_ID = A.X_ID
AND hist. X_VAL = A.X_VAL
)
sales_val - 50 lines mln (unique index)
sales_hist - 20 mln ranks (unique index)
I met many problems with waits and locks - how to optimize to avoid locks using the parameters of tables or changes of the declaration? or maybe is there another way to do optimization ?
Kind regards
Bolo
Thank you for that. Collect in bulk Unfortunatelly + FORALL does not work with the types in 10g - solution is to use the FOR loop and is not as effective as FORALL in many cases. I do still some tests to solve this problem - I'll put you once it's done.
EDIT: hash partitioning + fusion - 3-5 minutes (time), so for now it's the solution to the problem. Thank you all for the great discussion!
-
Handful of division by zero error in the update statement
My dear welcom
How and I overcame this problem in my update statement
for exwmple, it comes to the table
CREATE TABLE TEST4
(
A1 NUMBER,
A2 NUMBER OF,
A3 NUMBER
)
Insert into TEST4 a1, a2, values (15.3);
Insert into TEST4 a1, a2, values (9.3);
Insert into TEST4 a1, a2, values (12.0);
commit;
now, I'm trying to update the column a3 in Division a1/12
Start
update set of test4
A3 = a1/a2;
exception
WHEN ZERO_DIVIDE THEN
update set of test4
A3 = 0;
end;
After you validate THE VALUE of A3 = 0;
Can any onle help me
I use oracle 10g R2
Thanks in advanceSQL> select * 2 from test4 3 / A1 A2 A3 ---------- ---------- ---------- 15 3 9 3 12 0 SQL> update test4 2 set a3 = case a2 when 0 then 0 else a1 / a2 end 3 / 3 rows updated. SQL> select * 2 from test4 3 / A1 A2 A3 ---------- ---------- ---------- 15 3 5 9 3 3 12 0 0 SQL>SY.
-
How to use the escape character in the update statement.
Hi all
I'm trying to update using the following table update statement in sql, but whenever he asks me to enter the value 'and' below sql.
Please let me know how to use escape character syntax or let me know if there is no alternative.UPDATE xyz_xyz SET NAME = 'ABC & PQR' WHERE ID = (SELECT ID FROM abc_abc WHERE NAME = 'C & PQR');
Thank you
VishwasIf you use SQL * more then use SET DEFINE OFF.
Also, you can do something like this
UPDATE xyz_xyz SET NAME = 'ABC' ||'&' || 'PQR' WHERE ID = (SELECT ID FROM abc_abc WHERE NAME = 'C' ||'&' ||'PQR'); -
What happens all by issuing the UPDATE statement
Hi all
Can you please tell me, what happens when problem user update statement in sga with background process?
Thank you
Sanjeev.Hope you've finished reading your old post. "what happens all by issuing the select statement.
What happens all by issuing the select statementOpen the document since the last post and search "how Oracle database of the LMD processes."
-
Solution probably really Simple (where in the Update statement)
Right,
I have a table. In this table, there are 5 fields for the sake of arguments:
COLUMN_ONE COLUMN_THREE COLUMN_FOUR COLUMN_FIVE COLUMN_TWO
On this table there will be duplicate records, but only the column 1, column 2 and 3 on these "duplicate" files will be the same, column 4 will be different (column 4 is not a unique identifier, but if there are duplicate records, it will be different at least 2 records). Column 5 is not yet filled, all null.
I need a way to update column 5 of these duplicate records. However, let me update on one of the folders and the criteria for this will be where column_4 = "Hello".
So far I came with an update statement similar to the following, but I don't know where with her:
UPDATE DANS_TABLE
SET COLUMN_five = "DUPLICATE".
WHERE EXISTS (SELECT column_one | column_two | column_three, COUNT (*))
OF DANS_TABLE
GROUP BY column_one | column_two | column_three
SEEN (COUNT (*) > 1)) and column_four = 'Hello ';
For any help, thank you very much! Much appreciated,
DanThis is the query that updates that one line duplicated to column_five. I hope this helps...
UPDATE DANS_TABLE x 1
SET COLUMN_five = "DUPLICATE".
WHERE THERE ARE
(
Select r_id of
(select max (rowid) r_id, column_one | column_two | column_three)
of DANS_TABLE t1 where exists
(
Select str of
(
SELECT column_one | column_two | column_three str, COUNT (*)
OF DANS_TABLE
GROUP BY column_one | column_two | column_three
SEEN (COUNT (*) > 1)
) t2 where t2.str = t1.column_one | T1.column_two | T1.column_three
) column_one group | column_two | column_three
) x 2 where x 1 .rowid = 2 .r_id x
) and column_four = 'Hello '; -
Question about the update statement
Hi all
UPDATE table_name
SET Column1 = value, column2 = value2
WHERE condition;
I have an update like this statement. What happens when I run the above statement, when there is no column named Column2 in the table? If it creates a column named column2 and value2 using?
Instead should first create a new column named Column2 in the table first and then run this update statement?
Please, someone good to explain.Hi, this statement (it is a dml statement) will give an error, because the Column2 does not exist in the table. You must add the Column2 by an alter (ddl statement) statement. Then, you can try to execute your statement.
Kind regards
SabyPublished by: 793420 on September 7, 2010 14:48
Published by: 793420 on September 7, 2010 14:48
-
insertion of a comma in the records by using the update statement
Hello
I have a table with an address field LOCATIONS, and there are several files that contain Box or post office box before the street number, IE. "Box, 25 993 stone Ave", "1 194 John St box". What I would do, is add a comma between the box number and the street number for all records that start with "Box", IE 'box 25, Ave Pierre 993'. I have over 1000 addresses with a box number, an update statement would be more effective.
I created a display to select addresses that have a number of box with 1 digit (only addresses such as 'Box 1 194 St John'). I don't know how to write a statement that inserts a comma in the 6th space to create ' box 1, 194 John St.
Any ideas?You can use regular expressions
update place set address = regexp_replace(address, 'Box ([[:digit:]]+)', 'Box \1,') -
Performance issue with the Update statement
Oracle 10204
I have a problem related to updaing one performance table.
only 5000 lines should be updated.
Hir are some details on the tables/M.V concerned:
I tried to run the update with two options:TABLE_NAME LAST_ANALYZED NUM_ROWS SAMPLE_SIZE PS_RF_INST_PROD 1/20/2010 1:14:22 AM 7194402 7194402 BL_TMP_INTRNT 1/27/2010 9:08:54 AM 885445 885445 NAP_INTERNET 1/25/2010 11:47:02 AM 7053990 560777
1. with the plan than oracle choose.
2. with notes I added.
In both cases I he collapsed after more than an hour.
Can any one suggest how to speed it up?
Below are for the two option tkprof.
Plan 1Please note that beside the defualt statistics on those tables i also gathered statistics on two column level as followed: BEGIN SYS.DBMS_STATS.GATHER_TABLE_STATS ( OwnName => 'B' ,TabName => 'BL_TMP_INTRNT' ,Estimate_Percent => 100 ,Degree => 8 ,Cascade => TRUE ,No_Invalidate => FALSE); END; / exec dbms_stats.gather_table_stats('B' , 'BL_TMP_INTRNT', cascade=>TRUE, method_opt=>'for columns SERVICE_UID size 254'); exec dbms_stats.gather_table_stats('B' , 'BL_TMP_INTRNT', cascade=>TRUE, method_opt=>'for columns FIX_IP_USER size 254');
plan 2UPDATE BL_TMP_INTRNT A SET A.FIX_IP_USER = (SELECT C.PRODUCT_ID FROM NAP_INTERNET B, PS_RF_INST_PROD C WHERE B.INST_PROD_ID = A.SERVICE_UID AND B.SETID = 'SHARE' AND C.INST_PROD_ID = B.NAP_SURF_UID) WHERE A.TERM_DESC LIKE '%ip%' call count cpu elapsed disk query current rows ------- ------ -------- ---------- ---------- ---------- ---------- ---------- Parse 1 0.03 0.02 0 0 0 0 Execute 1 1166.64 4803.78 17989002 18792167 117 0 Fetch 0 0.00 0.00 0 0 0 0 ------- ------ -------- ---------- ---------- ---------- ---------- ---------- total 2 1166.67 4803.81 17989002 18792167 117 0 Misses in library cache during parse: 1 Optimizer mode: FIRST_ROWS Parsing user id: 13 Rows Row Source Operation ------- --------------------------------------------------- 0 UPDATE BL_TMP_INTRNT (cr=0 pr=0 pw=0 time=2 us) 46 TABLE ACCESS FULL BL_TMP_INTRNT (cr=586400 pr=22228 pw=0 time=15333652 us) 15 HASH JOIN (cr=18170453 pr=17931639 pw=0 time=3991064192 us) 46 MAT_VIEW ACCESS FULL NAP_INTERNET (cr=5659886 pr=5655436 pw=0 time=988162557 us) 329499624 MAT_VIEW ACCESS FULL PS_RF_INST_PROD (cr=12545734 pr=12311281 pw=0 time=2636471644 us)
UPDATE BL_TMP_INTRNT A SET A.FIX_IP_USER = (SELECT /*+ index(b NAP_INTERNET_PK) index(c PS_RF_INST_PROD_PK)*/ C.PRODUCT_ID FROM NAP_INTERNET B, PS_RF_INST_PROD C WHERE B.INST_PROD_ID = A.SERVICE_UID AND B.SETID = 'SHARE' AND C.INST_PROD_ID = B.NAP_SURF_UID) WHERE A.TERM_DESC LIKE '%ip%' call count cpu elapsed disk query current rows ------- ------ -------- ---------- ---------- ---------- ---------- ---------- Parse 1 0.02 0.02 0 0 0 0 Execute 1 4645.25 4613.70 95783 39798095 735 0 Fetch 0 0.00 0.00 0 0 0 0 ------- ------ -------- ---------- ---------- ---------- ---------- ---------- total 2 4645.27 4613.73 95783 39798095 735 0 Misses in library cache during parse: 1 Optimizer mode: FIRST_ROWS Parsing user id: 13 Rows Row Source Operation ------- --------------------------------------------------- 0 UPDATE BL_TMP_INTRNT (cr=0 pr=0 pw=0 time=1 us) 473 TABLE ACCESS FULL BL_TMP_INTRNT (cr=10461 pr=10399 pw=0 time=4629385 us) 408 MAT_VIEW ACCESS BY INDEX ROWID PS_RF_INST_PROD (cr=39776109 pr=85381 pw=0 time=4605125045 us) 1350 NESTED LOOPS (cr=39784584 pr=84974 pw=0 time=4601874262 us) 470 MAT_VIEW ACCESS BY INDEX ROWID NAP_INTERNET (cr=23569112 pr=50472 pw=0 time=2544364336 us) 470 INDEX FULL SCAN NAP_INTERNET_PK (cr=23568642 pr=50005 pw=0 time=2540300981 us)(object id 11027362) 408 INDEX FULL SCAN PS_RF_INST_PROD_PK (cr=16215472 pr=34502 pw=0 time=2057500175 us)(object id 10980137) Elapsed times include waiting on following events: Event waited on Times Max. Wait Total Waited ---------------------------------------- Waited ---------- ------------ db file scattered read 1300 0.50 4.27 db file sequential read 85707 0.51 29.88 latch: cache buffers chains 1 0.00 0.00 log file sync 1 0.00 0.00 SQL*Net break/reset to client 1 0.00 0.00 SQL*Net message to client 1 0.00 0.00 SQL*Net message from client 1 14.73 14.73 ********************************************************************************The problem in your update statement that is the query in your set clause is executed many times that there are lines in BL_TMP_INTRNT of 'intellectual property' in their column of term_desc. You mentioned there are about 5000, then the query joining NAP_INTERNET with PS_RF_ISNT_PROD is begun 5000 times.
The trick is to join only once, be updated using join views - provided that it is preserved - key or by using the merge statement.Here is an example:
SQL> create table bl_tmp_intrnt (fix_ip_user,service_uid,term_desc) 2 as 3 select level 4 , level 5 , 'aipa' 6 from dual 7 connect by level <= 5000 8 / Tabel is aangemaakt. SQL> create table nap_internet (inst_prod_id,setid,nap_surf_uid) 2 as 3 select level 4 , 'SHARE' 5 , level 6 from dual 7 connect by level <= 10 8 / Tabel is aangemaakt. SQL> create table ps_rf_inst_prod (product_id,inst_prod_id) 2 as 3 select level 4 , level 5 from dual 6 connect by level <= 10 7 / Tabel is aangemaakt. SQL> exec dbms_stats.gather_table_stats(user,'bl_tmp_intrnt') PL/SQL-procedure is geslaagd. SQL> exec dbms_stats.gather_table_stats(user,'nap_internet') PL/SQL-procedure is geslaagd. SQL> exec dbms_stats.gather_table_stats(user,'ps_rf_inst_prod') PL/SQL-procedure is geslaagd. SQL> set serveroutput off SQL> update ( select a.fix_ip_user 2 , c.product_id 3 from bl_tmp_intrnt a 4 , nap_internet b 5 , ps_rf_inst_prod c 6 where a.term_desc like '%ip%' 7 and a.service_uid = b.inst_prod_id 8 and b.setid = 'SHARE' 9 and b.nap_surf_uid = c.inst_prod_id 10 ) 11 set fix_ip_user = product_id 12 / set fix_ip_user = product_id * FOUT in regel 11: .ORA-01779: cannot modify a column which maps to a non key-preserved tableJoin is one key kept in the case of b.inst_prod_id and c.inst_prod_id are unique. Please refer to the documentation for more information here: http://download.oracle.com/docs/cd/B19306_01/server.102/b14231/views.htm#sthref3055
SQL> alter table nap_internet add primary key (inst_prod_id) 2 / Tabel is gewijzigd. SQL> alter table ps_rf_inst_prod add primary key (inst_prod_id) 2 / Tabel is gewijzigd. SQL> update /*+ gather_plan_statistics */ 2 ( select a.fix_ip_user 3 , c.product_id 4 from bl_tmp_intrnt a 5 , nap_internet b 6 , ps_rf_inst_prod c 7 where a.term_desc like '%ip%' 8 and a.service_uid = b.inst_prod_id 9 and b.setid = 'SHARE' 10 and b.nap_surf_uid = c.inst_prod_id 11 ) 12 set fix_ip_user = product_id 13 / 10 rijen zijn bijgewerkt. SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last')) 2 / PLAN_TABLE_OUTPUT -------------------------------------------------------------------------------------------------------------------------------------- SQL_ID c7nqbxwzpyq5p, child number 0 ------------------------------------- update /*+ gather_plan_statistics */ ( select a.fix_ip_user , c.product_id from bl_tmp_intrnt a , nap_internet b , ps_rf_inst_prod c where a.term_desc like '%ip%' and a.service_uid = b.inst_prod_id and b.setid = 'SHARE' and b.nap_surf_uid = c.inst_prod_id ) set fix_ip_user = product_id Plan hash value: 1745632149 --------------------------------------------------------------------------------------------------------------------------------------- | Id | Operation | Name | Starts | E-Rows | A-Rows | A-Time | Buffers | OMem | 1Mem | Used-Mem | --------------------------------------------------------------------------------------------------------------------------------------- | 1 | UPDATE | BL_TMP_INTRNT | 1 | | 0 |00:00:00.01 | 32 | | | | | 2 | NESTED LOOPS | | 1 | 10 | 10 |00:00:00.01 | 21 | | | | | 3 | MERGE JOIN | | 1 | 10 | 10 |00:00:00.01 | 9 | | | | |* 4 | TABLE ACCESS BY INDEX ROWID| NAP_INTERNET | 1 | 10 | 10 |00:00:00.01 | 2 | | | | | 5 | INDEX FULL SCAN | SYS_C00132995 | 1 | 10 | 10 |00:00:00.01 | 1 | | | | |* 6 | SORT JOIN | | 10 | 250 | 10 |00:00:00.01 | 7 | 267K| 256K| 237K (0)| |* 7 | TABLE ACCESS FULL | BL_TMP_INTRNT | 1 | 250 | 5000 |00:00:00.01 | 7 | | | | | 8 | TABLE ACCESS BY INDEX ROWID | PS_RF_INST_PROD | 10 | 1 | 10 |00:00:00.01 | 12 | | | | |* 9 | INDEX UNIQUE SCAN | SYS_C00132996 | 10 | 1 | 10 |00:00:00.01 | 2 | | | | --------------------------------------------------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 4 - filter("B"."SETID"='SHARE') 6 - access("A"."SERVICE_UID"="B"."INST_PROD_ID") filter("A"."SERVICE_UID"="B"."INST_PROD_ID") 7 - filter("A"."TERM_DESC" LIKE '%ip%') 9 - access("B"."NAP_SURF_UID"="C"."INST_PROD_ID") 32 rijen zijn geselecteerd. SQL> rollback 2 / Rollback is voltooid.And it's your current statement. Please note the number of 5000 in the column begins:
SQL> UPDATE /*+ gather_plan_statistics */ BL_TMP_INTRNT A 2 SET A.FIX_IP_USER = 3 (SELECT C.PRODUCT_ID 4 FROM NAP_INTERNET B, PS_RF_INST_PROD C 5 WHERE B.INST_PROD_ID = A.SERVICE_UID 6 AND B.SETID = 'SHARE' 7 AND C.INST_PROD_ID = B.NAP_SURF_UID) 8 WHERE A.TERM_DESC LIKE '%ip%' 9 / 5000 rijen zijn bijgewerkt. SQL> select * from table(dbms_xplan.display_cursor(null,null,'allstats last')) 2 / PLAN_TABLE_OUTPUT -------------------------------------------------------------------------------------------------------------------------------------- SQL_ID f1qtnpa0nmbh8, child number 0 ------------------------------------- UPDATE /*+ gather_plan_statistics */ BL_TMP_INTRNT A SET A.FIX_IP_USER = (SELECT C.PRODUCT_ID FROM NAP_INTERNET B, PS_RF_INST_PROD C WHERE B.INST_PROD_ID = A.SERVICE_UID AND B.SETID = 'SHARE' AND C.INST_PROD_ID = B.NAP_SURF_UID) WHERE A.TERM_DESC LIKE '%ip%' Plan hash value: 3190675455 ----------------------------------------------------------------------------------------------------------- | Id | Operation | Name | Starts | E-Rows | A-Rows | A-Time | Buffers | ----------------------------------------------------------------------------------------------------------- | 1 | UPDATE | BL_TMP_INTRNT | 1 | | 0 |00:00:00.10 | 5076 | |* 2 | TABLE ACCESS FULL | BL_TMP_INTRNT | 1 | 250 | 5000 |00:00:00.01 | 7 | | 3 | NESTED LOOPS | | 5000 | 1 | 10 |00:00:00.02 | 24 | |* 4 | TABLE ACCESS BY INDEX ROWID| NAP_INTERNET | 5000 | 1 | 10 |00:00:00.01 | 12 | |* 5 | INDEX UNIQUE SCAN | SYS_C00132995 | 5000 | 1 | 10 |00:00:00.01 | 2 | | 6 | TABLE ACCESS BY INDEX ROWID| PS_RF_INST_PROD | 10 | 10 | 10 |00:00:00.01 | 12 | |* 7 | INDEX UNIQUE SCAN | SYS_C00132996 | 10 | 1 | 10 |00:00:00.01 | 2 | ----------------------------------------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 2 - filter("A"."TERM_DESC" LIKE '%ip%') 4 - filter("B"."SETID"='SHARE') 5 - access("B"."INST_PROD_ID"=:B1) 7 - access("C"."INST_PROD_ID"="B"."NAP_SURF_UID") 29 rijen zijn geselecteerd.Kind regards
Rob.
Maybe you are looking for
-
Divide the 13 x 2 PC: mode 'SLEEP'
Hello I am owner of a Split 13 x 2 PC and I facing the following problem. When I close the lid of the portable computer will not sleep and the fan still works. I tried a different function for the lid and instead of mode 'sleep', I chose the closure
-
Hello. I do not understand why my Mac find see my iMac, but is unable to see its location? + A sound was played on my iMac to my iPhone when I test the sound function. Please see the pictures on my preferences. PS: My iMac connected life ethernet no
-
Satellite Pro L550-17U fails to start, disk noise only
Hello I have a Toshiba Satellite Pro L550-17U age of 3 years (Sept 2010). And it worked perfectly ok since I bought it. 2 weeks ago I turned it on, my coffee table, where 90% of the time and nothing! Black screen, no keyboard lights (Caps/screen/NumL
-
Late in the evolution of the status of the LEDs
Hello I use 6 LEDs in my application. I change the status of the LED using the SetCtrlVal() function. But for 3 LEDS, change happens with a certain delay (delay of about 5 sec). I do not know how to solve in this issue. All me please on this. Thanks
-
Replace my old data on previous installation of vista.
My pc no longer starts (Vista has not started). They advised me to reinstall vista, which I did, but no, I can't go back to my old data on the previous installation of vista. Any help how to recover this data?