Commit in every statement?

Hi all,
why is it bad to place "commit;" in the end of each "insert into" statement? What should I know before doing that?

Apart fromthe fact that a commit should be done at the end of a logical/business transaction, there are technical reasons not to do it too.
In simplistic terms...
If you look at the underlying processes on the database server, there are things called "writer" processes. By default Oracle has a few of these going at any one time and they handle the need to take date when you commit and write it to the underlying o/s files. These write processes run in parallel to the database server engine, so they get on with the job of writing the data as and when they can whilst the database engine gets on with what it needs to do, thus your applications don't have to wait for data to actually be physically written to the disks before continuing.
If you start issuing lots of update/insert statements with lots of commits, then the existing writer processes get overloaded with work to do (as they are required to treat each committed transaction individually), and the oracle database process ends up spawning more writer processes to handle the additional workload. The more processes that get started, all trying to write to the disks as the same time, the slower the database server will run, and the slower your application at the front end will appear to go.
"Commit" is like saying "I need this data written to the disks as soon as possible" (and thus available to other sessions). In truth there not often a need to commit so often and really have the data there a.s.a.p. and this is where we get the priniciple that you should only commit when it is logical to do so from a technical or business need, rather than just when you feel like it.

Similar Messages

Commit for every 1000 records in Insert into select statment

Hi I've the following INSERT into SELECT statement .
The SELECT statement (which has joins ) has around 6 crores fo data . I need to insert that data into another table.
Please suggest me the best way to do that .
I'm using the INSERT into SELECT statement , but i want to use commit statement for every 1000 records .
How can i achieve this ..
insert into emp_dept_master
select e.ename ,d.dname ,e.empno ,e.empno ,e.sal
   from emp e , dept d
where e.deptno = d.deptno       ------ how to use commit for every 1000 records .Thanks

Smile wrote:
Hi I've the following INSERT into SELECT statement .
The SELECT statement (which has joins ) has around 6 crores fo data . I need to insert that data into another table.Does the another table already have records or its empty?
If its empty then you can drop it and create it as
create your_another_table
as
<your select statement that return 60000000 records>
Please suggest me the best way to do that .
I'm using the INSERT into SELECT statement , but i want to use commit statement for every 1000 records .That is not the best way. Frequent commit may lead to ORA-1555 error
[url http://asktom.oracle.com/pls/apex/f?p=100:11:0::::P11_QUESTION_ID:275215756923]A nice artical from ASKTOM on this one
How can i achieve this ..
insert into emp_dept_master
select e.ename ,d.dname ,e.empno ,e.empno ,e.sal
from emp e , dept d
where e.deptno = d.deptno       ------ how to use commit for every 1000 records .
It depends on the reason behind you wanting to split your transaction into small chunks. Most of the time there is no good reason for that.
If you are tying to imporve performance by doing so then you are wrong it will only degrade the performance.
To improve the performance you can use APPEND hint in insert, you can try PARALLEL DML and If you are in 11g and above you can use [url http://docs.oracle.com/cd/E11882_01/appdev.112/e25788/d_parallel_ex.htm#CHDIJACH]DBMS_PARALLEL_EXECUTE to break your insert into chunks and run it in parallel.
So if you can tell the actual objective we could offer some help.

Need to commit after every 10 000 records inserted ?

What would be the best way to Commit after every 10 000 records inserted from one table to the other using the following script :
DECLARE
l_max_repa_id x_received_p.repa_id%TYPE;
l_max_rept_id x_received_p_trans.rept_id%TYPE;
BEGIN
SELECT MAX (repa_id)
INTO l_max_repa_id
FROM x_received_p
WHERE repa_modifieddate <= ADD_MONTHS (SYSDATE, -6);
SELECT MAX (rept_id)
INTO l_max_rept_id
FROM x_received_p_trans
WHERE rept_repa_id = l_max_repa_id;
INSERT INTO x_p_requests_arch
SELECT *
FROM x_p_requests
WHERE pare_repa_id <= l_max_rept_id;
DELETE FROM x__requests
WHERE pare_repa_id <= l_max_rept_id;

1006377 wrote:
we are moving between 5 and 10 million records from the one table to the other table and it takes forever.
Please could you provide me with a script just to commit after every x amount of records ? :)I concur with the other responses.
Committing every N records will slow down the process, not speed it up.
The fastest way to move your data (and 10 million rows is nothing, we do those sorts of volumes frequently ourselves) is to use a single SQL statement to do an INSERT ... SELECT ... statement (or a CREATE TABLE ... AS SELECT ... statement as appropriate).
If those SQL statements are running slowly then you need to look at what's causing the performance issue of the SELECT statement, and tackle that issue, which may be a case of simply getting the database statistics up to date, or applying a new index to a table etc. or re-writing the select statement to tackle the query in a different way.
So, deal with the cause of the performance issue, don't try and fudge your way around it, which will only create further problems.

Avoid Commit after every Insert that requires a SELECT

Hi everybody,
Here is the problem:
I have a table of generator alarms which is populated daily. On daily basis there are approximately 50,000 rows to be inserted in it.
Currently i have one month's data in it ... Approximately 900,000 rows.
here goes the main problem.
before each insert command, whole table is checked if the record does not exist already. Two columns "SiteName" and "OccuranceDate" are checked... this means, these two columns are making a unique record when checked together with an AND operation in WHERE clause.
we have also implemented partition on this table. and it is basically partitioned on the basis of OccuranceDate and each partition has 5 days' data.
say
01-Jun to 06 Jun
07-Jun to 11 Jun
12-Jun to 16 Jun
and so on
26-Jun to 30 Jun
NOW:
we have a commit command within the insertion loop, and the each row is committed once inserted, making approximately 50,000 commits daily.
Question:
Can we commit data after say each 500 inserted rows, but my real question is can we Query the records using SELECT which are Just Inserted but not yet committed ?
a friend told me that, you can query the records which are inserted in the same connection session but not yet committed.
Can any one help ?
Sorry for the long question but it was to make u understand the real issue. :(
Khalid Mehmood Awan
khalidmehmoodawan @ gmail.com
Edited by: user5394434 on Jun 30, 2009 11:28 PM

Don't worry about it - I just said that because the experts over there will help you much better. If you post your code details there they will give suggestions on optimizing it.
Doing a SELECT between every INSERT doesn't seem very natural to me, but it all depends on the details of your code.
Also, not committing on time may cause loss of the uncommitted changes. Depending on how critical the data is and the dependency of the changes, you have to commit after every INSERT, in between, or at the end.
Regards,
K.

Commit after every 1000 records

Hi dears ,
i have to update or insert arround 1 lakhs records every day incremental basis,
while doing it , after completing all the records commit happens, In case some problem in between all my processed records are getting rollbacked,
I need to commit it after every frequency of records say 1000 records.
Any one know how to do it??
Thanks in advance
Regards
Raja

Raja,
There is an option in the configuration of a mapping in which you can set the Commit Frequency. The Commit Frequency only applies to non-bulk mode mappings. Bulk mode mappings commit according the bulk size (which is also an configuration setting of the mapping).
When you set the Default Operating Mode to row based and Bulk Processing Code to false, Warehouse Builder uses the Commit Frequency parameter when executing the package. Warehouse Builder commits data to the database after processing the number of rows specified in this parameter.
If you set Bulk Processing Code to true, set the Commit Frequency equal to the Bulk Size. If the two values are different, Bulk Size overrides the commit frequency and Warehouse Builder implicitly performs a commit for every bulk size.
Regards,
Ilona

COMMIT after every 10000 rows

I'm getting probelms with the following procedure. Is there any that I can do to commit after every 10,000 rows of deletion? Or is there any other alternative! The DBAs are not willing to increase the undo tablespace value!
create or replace procedure delete_rows(v_days number)
is
l_sql_stmt varchar2(32767) := 'DELETE TABLE_NAME WHERE ROWID IN (SELECT ROWID FROM TABLE_NAME W
where_cond VARCHAR2(32767);
begin
   where_cond := 'DATE_THRESHOLD < (sysdate - '|| v_days ||' )) ';
   l_sql_stmt := l_sql_stmt ||where_cond;
   IF v_days IS NOT NULL THEN
       EXECUTE IMMEDIATE l_sql_stmt;
   END IF;
end;I think I can use cursors and for every 10,000 %ROWCOUNT, I can commit, but even before posting the thread, I feel i will get bounces! ;-)
Please help me out in this!
Cheers
Sarma!

Hello
In the event that you can't persuede the DBA to configure the database properly, Why not just use rownum?
SQL> CREATE TABLE dt_test_delete AS SELECT object_id, object_name, last_ddl_time FROM dba_objects;
Table created.
SQL>
SQL> select count(*) from dt_test_delete WHERE last_ddl_time < SYSDATE - 100;
COUNT(*)
     35726
SQL>
SQL> DECLARE
2
3     ln_DelSize                      NUMBER := 10000;
4     ln_DelCount                     NUMBER;
5
6 BEGIN
7
8     LOOP
9
10             DELETE
11             FROM
12                     dt_test_delete
13             WHERE
14                     last_ddl_time < SYSDATE - 100
15             AND
16                     rownum <= ln_DelSize;
17
18             ln_DelCount := SQL%ROWCOUNT;
19
20             dbms_output.put_line(ln_DelCount);
21
22             EXIT WHEN ln_DelCount = 0;
23
24             COMMIT;
25
26     END LOOP;
27
28 END;
29 /
10000
10000
10000
5726
0
PL/SQL procedure successfully completed.
SQL>HTH
David
Message was edited by:
david_tyler

Same code for every state in event

Hello.
In my front panel there are some buttons that will give the same result each time they are pressed.
Exactly like menu items.
In every state of my FSM there is an event structure. Is it efficient to add a case for every menu item and buttons for every state?
Is there a way to add this functionality in just one case structure and then make it global somehow?
Thanks.

duplicate, continue here!
Best regards,
GerdW
CLAD, using 2009SP1 + LV2011SP1 + LV2014SP1 on WinXP+Win7+cRIO
Kudos are welcome

Is it possible to automatically add a comma after every 6 characters has been typed in a text field?

Hope someone can help me.
What I basically need help with is how to make Acrobat add a comma after every 6 characters in a text field:
XXXXXX,YYYYYY, ZZZZZZ etc

I'm sorry, but I did not understand that (i'm using Acrobat Pro X)
Am I supposed to go to:
Text Field Properties > Format > Custom
and then use Custom Format Script or Custom Keystroke Script?
I tried both and it did not work.
And do the Text Field have to be named "chunkSize"?
Seems like it works. I had to move to the next formfield in order to see the effect.
Is it possible to make it happen in real time (as you type the comma is inserted?)

Commit after every three UPDATEs in CURSOR FOR loop

DB Version: 11g
I know that experts in here despise the concept of COMMITing inside loop.
But most of the UPDATEs being fired by the code below are updating around 1 million records and it is breaking our UNDO tablespace.
begin
for rec in
      (select owner,table_name,column_name
      from dba_tab_cols where column_name like 'ABCD%' and owner = p_schema_name)
      loop
        begin
        execute immediate 'update '||rec.owner||'.'||rec.table_name||' set '||rec.column_name|| ' = '''||rec.owner||'';
        end;
      end loop;
end;We are not expecting ORA-01555 error as these are just batch updates.
I was thinking of implementing something like
FOR i IN 1..myarray.count
LOOP
                         DBMS_OUTPUT.PUT_LINE('event_key at' || i || ' is: ' || myarray(i));
                         INSERT INTO emp
                         empid,
                         event_id,
                         dept,
                         event_key
                         VALUES
                         v_empid,
                         3423,
                         p_dept,
                         myarray(i)
                         if(MOD(i, p_CommitFreq) = 0) --- When the loop counter becomes fully divisible by p_commit_frequency, it COMMITs
                         then
                                   commit;
                         end if;
END LOOP;(Found from an OTN thread)
But i don't know how to access the loop counter value in a CURSOR FOR loop.

To be fair, what is really despised is code that takes an operation that could have been performed in a single SQL statement and steps through it in the slowest possible way, committing pointlessly as it goes along (exactly like the example you found). Your original version doesn't do that - it looks more like some sort of one-off migration where you have to set every value of every column that matches some naming standard pattern to a constant. If that's the case, and if there are huge volumes involved and you can't simply add a bit more undo, then I don't see much wrong with committing after each update, especially if you track how far you've got so you can restart cleanly if it fails.
If you really want an incrementing counter in an unnamed cursor, apart from the explicit variable others have suggested, you could add rownum to the cursor (alias it to something that isn't an Oracle keyword), although it could complicate the ORDER BY that you might be considering for the restart logic. You could have that instead of the redundant 'owner' column in your example which is always the same as the constant p_schema_name.
Another approach would be to keep track of SQL%ROWCOUNT after each update by adding it to a variable, and commit when the total number of rows updated so far reaches, say, a million.
Could the generated statement use a WHERE clause, or does it really have to update every row in every table it finds?
Could there be more than one column to update per table? If so it might be worth generating one multi-column update statement per table, although it'll complicate things a bit.
btw you don't need the inner 'begin' and 'end' keywords, and whoever supplied the MOD example you found should know that three or four spaces usually make a good indent and you don't need brackets around IF conditions.

Commit to SQL statement

Hello N. Gasparotto,
I have written and tried the code.
I got the desired result.
Actually I have a form. That form has "slope" as label with the list of values:
0-2%
2-6%
10-15%
6-8% and so on.
the code for that form is written in jsp which includes sql statements.
whenever i make changes to the database through jsp, i generally commit the changes in the database so that i can see the changes in the form.
After execution of your code, i got the desired result.at the sql prompt, i typed commit. it displayed commit complete. The JSP code contains the sql statement:
"select item from code_slope where code_slope<>'9999' order by 2 asc;"
So after execution of this jsp, I would like to see the list of values in the following order:
Slope:
0-2%
2-6%
6-8%
10-15%.
But I couldn't find the change in the form.
Could you guide me what i need to do next when i execute format_numeric function and the other code so that i could see the change in the form?
Waiting for reply!
Prathima.
Waiting for reply!

Hello sir,
I am unable to express my problem. Let me explain what my problem is with an example that best suits my current problem.
For example, I have a jsp page with field "Level".That jsp page contains code(for example, select level from tname order by 2 asc) for that field "Level".If we assume that the field Level has a list box with values "One, Two and so on". Now on the database side I would like to add new value into that Level field "Three".After inserting new value into that field, I commit the change on the database side(sql>commit). When I refresh the jsp page, I could see the new value in the existing list box values.
Coming to the current problem, when I execute format_numeric and sort function, i could get the desired output. But when i refresh that particular jsp, i could not find the values in the sorted order.Could you please let me know what i need to do when i execute the format_numeric function + sort statement.
Eagerly Waiting for reply!
Prathima

Why is OEM SQL Monitoring showing parallel on almost every statement

I'm confused here.
I'm running Oracle EE 11.2.0.2 and when I look in OEM SQL Monitoring, it shows nearly every sql statement running with a degree of parallelism of *2*.
I've checked dba_tables and the 'degree' for all tables is only 1.
I look at the actual sql statement, and there are no hints to tell it to use parallelism.
So why and how is the database using parallelism?
I do see that parallel_threads_per_cpu is set to 2, but this is default for our Solaris 10 operating system.
REF: (for 11.2)
===========
PARALLEL_THREADS_PER_CPU specifies the default degree of parallelism for the instance and determines the parallel adaptive and load balancing algorithms. The parameter describes the number of parallel execution processes or threads that a CPU can handle during parallel execution.
The default is platform-dependent and is adequate in most cases. You should decrease the value of this parameter if the machine appears to be overloaded when a representative parallel query is executed. You should increase the value if the system is I/O bound.
I guess the next question here is how to tell if my database is actually IO bound or not?

Hi John. Thanks for your reply.
NAME TYPE VALUE
parallel_degree_policy string MANUALAnd so, the more I read about PARALLEL_THREADS_PER_CPU, the more I wonder if I should increase this value.
But first, I want to understand why I'm seeing parallelism in OEM set to 2 for almost everything that runs in the database, but note, NOT ALL.
Some queries, especially those running from Crystal Reports, are not using parallelism at all.
Is it possible to set a parameter at the session level that runs parallelism, and perhaps this is being done by the application?
I'm going to try increasing my PARALLEL_THREADS_PER_CPU to 4 and see if this changes the parallelism in OEM, (but I doubt it).
I should note that my most recent AWR report shows my db_file_sequential_read in the top 5 wait events.
This would imply my index reads and table reads by ROWID are waiting for disk - possibly I/O bound.
Edited by: 974632 on Jan 28, 2013 10:25 AM

How to make a US map with interactive buttons for every state ...

I have a photoshop map of the unitedstates with the states seperated. what I would like to do is put this map in dreamweaver so that each state is a button....any ideas?

imagemap
best,
Shocker

How do I read DAQ at every state change on a quad encoder over RTSI

I have successfully routed either A or B encoder phase to the DAQ card over RTSI, but this gives you 1/4 of your encoder counts to the DAQ. Is there a way to trigger the DAQ clock with both the rising and falling edges, A and B signals so you get a DAQ reading with every encoder count?

mikema111,
From your explanation I am assuming that you are using an X4 encoder. Unfortunately there isn't a way to combine both Phase A and Phase B (rising and falling edges) into a DAQ scan clock without external circuitry.
However, one possibility would be to use a sort of XOR circuit to merge the two phases into one signal and then pass that signal into one of your analog input channels. You could then setup that analog input channel for windowed analog input triggering. As the TTL pulse rises through the window an AI Start Trigger pulse is generated and then another pulse is generated as it passes back down through this window. Pass the AI Start Trigger pulse into a counter setup for retriggerable single pulse generation and you will have your DAQ scan cloc
k on the output pin of the counter.
If you are just interested in counting the pulses in both Phase A and Phase B you can configure one counter to count on the rising edge and the other on the falling edge as described in the following Knowledge Base:
http://digital.ni.com/public.nsf/websearch/15170E05F0F4B65C86256E2400812CD9?OpenDocument
I also recommend reading the following document that discusses several different options when using a quadrature encoder with an E Series board (DAQ-STC).
http://zone.ni.com/devzone/devzoneweb.nsf/Opendoc?openagent&36BD71244BB26FC886256869005E541B
Ames
Applications Engineering
National Instruments

Insert comma at every third position in string...

I am flumoxed on something rather simple I fear. I have a string like this ...
B01B09B20B21C13E10F07G12G20G24
I need to turn it into this
B01,B09,B20,B21,C13,E10,F07,G12,G20,G24
It is consistant in the following:
1. each segment will always be 3 characters long
2. each segment will always be structured as 1 character and 2 numerals
3. the list will always vary in length but always divisible by 3
Any simple solutions? I have though about various cflooping methods and simply not liked anythign I cam up with.
All help is greatly appreciated.
God Bless!
Chris

Here are a couple options. I prefer the first.
<cfset string = "B01B09B20B21C13E10F07G12G20G24" />
<cfset newstring = "" />
<cfloop from="1" to="#len(string)#" index="i" step="3">
<cfset newstring = listAppend(newstring, mid(string, i, 3), ",") />
</cfloop>
<cfoutput>#newstring#</cfoutput>
========================================================================================== =
<cfset string = "B01B09B20B21C13E10F07G12G20G24" />
<cfset counter = 0 />

<cfloop from="0" to="#len(string)-4#" index="i" step="3">
<cfset string = insert(",", string, i+counter+3) />
<cfset counter++ />
</cfloop>
<cfoutput>#string#</cfoutput>

JDBC receiver - INSERT issue

Hi all.
Let's assume I have a document with a number of INSERT statements :
- <p2:XXX xmlns:p2="http://abc.com">
- <UPDATE_INSERT>
- <IN_MAT action="INSERT">
<table>MY_TAB</table>
- <access>
<MANDT>030</MANDT>
<WERK>LW01</WERK>
<MATNR>PFS</MATNR>
<INTIME>04.07.2005</INTIME>
<MATTYP>0012</MATTYP>
<MATXT>PFS</MATXT>
<GMIEN>PJM</GMIEN>
<RATIO>0000000001.000</RATIO>
</access>
</IN_MAT>
</UPDATE_INSERT>
- <UPDATE_INSERT>
- <IN_MAT action="INSERT">
<table>MY_TAB</table>
- <access>
<MANDT>030</MANDT>
<WERK>LG01</WERK>
<MATNR>HPL</MATNR>
<INTIME>16.06.2005</INTIME>
<MATTYP>0013</MATTYP>
<MATXT>HPL</MATXT>
<GMIEN>ARK</GMIEN>
<RATIO>0000000005.330</RATIO>
</access>
</IN_MAT>
</UPDATE_INSERT>
Is there any way to process them in separate way. I mean
that in case of failure (for any reason) in one INSERT
rest of them wasn't rolled back??
I guess, that usually a whole XML message is rolled back.
Am I right?
Regards,
Grzegorz.

Just a wild option...:) ..
i think it will work...but it may not be a very pretty one...Just thinking along the lines of having a commit after every statement...Your message would be something like this...
- <p2:XXX xmlns:p2="http://abc.com">
- <UPDATE_INSERT>
- <IN_MAT action="INSERT">
<table>MY_TAB</table>
- <access>
<MANDT>030</MANDT>
<WERK>LW01</WERK>
<MATNR>PFS</MATNR>
<INTIME>04.07.2005</INTIME>
<MATTYP>0012</MATTYP>
<MATXT>PFS</MATXT>
<GMIEN>PJM</GMIEN>
<RATIO>0000000001.000</RATIO>
</access>
</IN_MAT>
</UPDATE_INSERT>
<Commit_STAT>
<docommit action="SQL_DML">
<access>COMMIT WORK</access>
</docommit>
</Commit_STAT>
- <UPDATE_INSERT>
- <IN_MAT action="INSERT">
<table>MY_TAB</table>
- <access>
<MANDT>030</MANDT>
<WERK>LG01</WERK>
<MATNR>HPL</MATNR>
<INTIME>16.06.2005</INTIME>
<MATTYP>0013</MATTYP>
<MATXT>HPL</MATXT>
<GMIEN>ARK</GMIEN>
<RATIO>0000000005.330</RATIO>
</access>
</IN_MAT>
</UPDATE_INSERT>
<Commit_STAT>
<docommit action="SQL_DML">
<access>COMMIT WORK</access>
</docommit>
</Commit_STAT>
Thanks & Regards,
Renjith

Commit in every statement?

Similar Messages

Maybe you are looking for