Row level locks

Hi All,
Version - Oracle Database 10g Enterprise Edition Release 10.2.0.2.0
I have a card's data in a table. Whenver a two different sessions/threads selects the data sometimes they get the same card.
Now I need to lock the record. If Thred-1 selects then it will lock the first records and thread-2 must select the second record as the first record is locked by thread-1.
I have written a piece of code , at first session it locks one record and when second session select it selects no records.
I have takem MIN group function because I want the cards to be selected in order.
Can you please tell where is my code went wrong or need some changes. Data and code is given below.
SQL> SELECT * FROM crm_pps_cards_sz ORDER BY 1;
CARD_NO          PIN SERIAL_NUMBER DATE_CREATED DATE_MERGED DATE_ALLOCATED DATE_REGISTERED
6338079966430591 9985 9950013661    12/06/2011
6338079973369543 6858 9950013660    12/06/2011
6338079978994154 7144 9950013655    12/06/2011
6338079981471778 7631 9950013654    12/06/2011
6338079986365041 7849 9950013657    12/06/2011
SQL>
Thread - 1
Connected to Oracle Database 10g Enterprise Edition Release 10.2.0.2.0
Connected as crm
SQL> set serveroutput on
SQL>
SQL> DECLARE
2    l_rec crm_pps_cards_sz%ROWTYPE;
3    CURSOR c IS
4      SELECT * FROM crm_pps_cards_sz t WHERE t.card_no =
5      (SELECT MIN(card_no) card FROM crm_pps_cards_sz)
6      FOR UPDATE NOWAIT SKIP LOCKED;
7
8 BEGIN
9    OPEN c;
10    FETCH c
11      INTO l_rec;
12    CLOSE c;
13    dbms_output.put_line('Card Allocated ' || l_rec.card_no);
14
15 END;
16 /
Card Allocated 6338079966430591
PL/SQL procedure successfully completed
SQL>
Thread - 2
Connected to Oracle Database 10g Enterprise Edition Release 10.2.0.2.0
Connected as crm
SQL> set serveroutput on
SQL>
SQL> DECLARE
2    l_rec crm_pps_cards_sz%ROWTYPE;
3    CURSOR c IS
4      SELECT * FROM crm_pps_cards_sz t WHERE t.card_no =
5      (SELECT MIN(card_no) card FROM crm_pps_cards_sz)
6      FOR UPDATE NOWAIT SKIP LOCKED;
7
8 BEGIN
9    OPEN c;
10    FETCH c
11      INTO l_rec;
12    CLOSE c;
13    dbms_output.put_line('Card Allocated ' || l_rec.card_no);
14
15 END;
16 /
Card Allocated
PL/SQL procedure successfully completed
SQL> My concern is that it must allocate in the order
6338079966430591
6338079973369543
....

It is because MIN(ard_no) is able to find value from table which is locked by you thread 1 and thread 1 mentioned NOWAIT thus it thread 2 went out without getting anything.
Please check my reply to your duplicate thread "row level locking" .

Similar Messages

Pl sql row level locking with wait ?

Hi,
I am using oracle 10g.
I am new to Oracle locks. I have two tables Table1(id_no, employee, salary) and Table2(id_no, employee, salary).
I need to pull any requested row from Table1 into Table2 only once on demand. I have a procedure to pull data and there could be more than one requests try to call same procedure to pull a row from Table1 into Table2 at any given time.
I coded below to achieve row level lock. if one transaction gets row level lock on Table1 at 2, so other Transactions should wait till the lock is released at line 2 or 5 to avoid duplicates.
But below code is not working, I am getting duplicates when I call this using two concurrent java threads.
How do I control this concurrency issue so that I can avoid duplicate entries in Table2. Could any one please help?
1.begin
2.select 0 into emp_cnt
3.from Table1 where id=id_no
4.for update;
5.update Table1 set employee='xyz'
6.where id=id_no;
7.select count(*) into table2_cnt from Table2 where id=id_no;
8.if(table2_cnt =0) then
9.code to insert above row from Table1 to Table2;
10.end if;
11.commit;
12.End;
Edited by: 980916 on Jan 9, 2013 5:48 PM

Welcome to the forums and welcome to Oracle.
Lets establish one thing right from the beginning ... the Oracle Database is not a Microsoft product.
There is no general reason to use row level locks, you should not want to use row level locks, you don't need row level locks, and you will almost never have any valid reason to consider row level locks.
That said there are two situations were it may be necessary to lock a row prior to an update or delete (possibly in a merge) and in those cases you want to use the built-in SELECT FOR UPDATE syntax. (Demo here: http://www.morganslibrary.org/reference/deadlocks.html#dlfu). But we should be clear here ... one rarely needs to use this locking mechanism as the chances of a collision in a well designed application are essentially zero.
Also as you are new to Oracle please explore the dynamic performance view V$RESERVED_WORDS and do not name objects, columns, etc. with reserved word names not that there is an excuse in any product to name a column "ID." Something Joe Celko has railed about for decades and a tradition I think we should all follow.
In the case of your posted code example (btw please read the FAQ and learn how to use tags) the solution is SELECT FOR UPDATE if locking can be justified which is unlikely.
Edited by: damorgan on Jan 9, 2013 6:11 PM

Detail screen for row level locking

Hi All ,
I have developed row level locking for a table , Next requirment is to display the details of each row i another screen when the user click on each row .And if the row which is clicked for display is geyout in first screen then it should be grey in details screen also .If the row is editable in the first screen then it should be in change mode in details screen .
I appreciate the help and support .
please provide the answer as soon as possbile .....
Thanks,
NJ

Yes, if a process holds many row level locks on a table, those locks by default can be escalated to partition locks, and, if necessary, table locks. Unless you are getting and holding many locks on a given table, this will not happen.
You say you are exporting data. If all you are doing is selecting from the table (not updating, deleting or inserting data) and you are using the read committed transaction isolation level, you would only hold a lock on a row for the small fraction
of a second when you are actually reading it. So you will never have large numbers of locks at one time. If you are doing large numbers of updates, inserts, or deletes in a singletransaction or if you are doing large numbers of selects in a single
transaction with the isolation level set to repeatable read or serializable, then it is possible to get enough locks to cause SQL server to escalate to partition or table locks. This is normally a good thing (that's the reason Microsoft made it the default).
To many locks eats up memory.
However, if you determine this is causing you a problem and you do not want the locks on a particular table ever escalated, you can use the ALTER TABLE SET (LOCK_ESCALATION = DISABLE) command to prevent most cases of lock escalation on a table by table basis.
See BOL for a description of the ALTER TABLE command. However, I would recommend you not use this unless you have determined that lock escalation is causing you problems. It is very easy to make things worse, not better by doing this.
Tom

Row Level Locking while inserting a record.

It is a good practice to lock the whole table or do the row level locking while performing any DELETE / UPDATE / MODIFY actions on a database table. Is it necessary to do the same thing while inserting a record via INSERT statement?
One point may arise if two users are inserting same records at a same time....
Well i am little bit confused here bcos if a record doen't exist in a table what is the point of locking it.
Please help me.

create a lock object using SE11 for that perticular table and include field names in Lock parameters. Then it will generate two FMs one for locking and another for unlocking.
Call the lock FM before updating the table and pass that row key value(For fields which taken in lock parameters for creating lock objects) to the exporting parameters.
Then do the updation.
Reward if useful...................

GTT table getting Row Exclusive lock

I have a procedure which loads a table which is used for reporting.
Currently it is a single process which picks up most of the data after joining 2-3 big tables(around 70-80GB) and then loading the main table.
The joins are on PI and also partitions are also being used. Then a lot of updates are done on the main table to update other columns which are being picked from different tables.
As the processing is happening on huge amount of data(around 1M), so processing is taking a lot of time.
This process is taking around 40-45 minutes.
I am planning to use parallelism to run it in 75 sessions. So the initial big insert will be faster and later on all the updates will be done on smaller chunks which will benefit the performance.
I planned to use GTT table so that i dont have to create 75 temp tables for each sessions.
But while using GTT table(ON COMMIT DELETE ROWS) in parallel sessions, i am seeing that the GTT table is getting Row Exclusive lock and the time taken by overall process is increasing by lot.
So i tested by using 75 temp tables. There i saw that the performance has increased by lots as assumed initially.
Please let me know if there is any other way or how to remove the locks on GTT table.

First you should question why you think you need 75 GTT.
Also your question contains no useful information
(no four digit version of Oracle, no OS, no number of processors) , this paragrah
Currently it is a single process which picks up most of the data after joining 2-3 big tables(around 70-80GB) and then loading the main table.
The joins are on PI and also partitions are also being used. Then a lot of updates are done on the main table to update other columns which are being picked from different tables.
As the processing is happening on huge amount of data(around 1M), so processing is taking a lot of time.
This process is taking around 40-45 minutes.tells nothing,
so basically your questions boils down to
- Hey I come from a sqlserver background (this is just an educated guess), how can I find a workaround to wreck Oracle to make it work like sqlserver.
75 parallel sessions on say 4 to 8 processors is a very bad idea, as these sessions will be simple competing for resources.
Also a row exclusive lock is just that: a row level lock. This isn't a problem, usually.
Sybrand Bakker
Senior Oracle DBA

How to get a row level share lock

Is it possible to acquire a row-level share lock, which would do all of the following?
1. Prevent others from updating that row.
2. Allow others to read that row.
3. Allow others to update other rows in the same table.
I have the following scenario where two transactions need to lock eachother out:
Set-up:
Insert into TABLE_A(value_a) values ('ok');
Insert into TABLE_B(value_b) values ('ok');
Transaction A:
Select value_b from TABLE_B
If value_b = 'ok', update TABLE_A set value_a = 'not ok'
Transaction B:
Select value_a from TABLE_A
If value_a = 'ok', update TABLE_B set value_b = 'not ok'
If transaction A runs first then the end result is "not ok" only in TABLE_A.
If transaction B runs first then the end result is "not ok" only in TABLE_B.
If the two transactions run concurrently, it is possible to get "not ok" in both tables. This is what I would like to prevent.
One way to get what I want is to use "select for update":
Transaction A:
Select value_a from TABLE_A for update
Select value_b from TABLE_B for update
If value_b = 'ok', update TABLE_A set value_a = 'not ok'
Transaction B:
Select value_a from TABLE_A for update
Select value_b from TABLE_B for update
If value_b = 'ok', update TABLE_B set value_a = 'not ok'
This way both transactions won't perform their update unless they know that the result if their select will still be the same after they commit. However, by using "select for update" Transaction A has gained an exclusive lock on the TABLE_B row. If a Transaction C with the same contents as Transaction A happens concurrently, then the two will block eachother even though all they both want is to read data from the same table.
Another way is to use "lock table", however using that would block out not only writes to a specific row, but writes to all rows in the table. (In my example there is only one row, but obviously that's just a simplified example.)
I have looked at the "serializable" isolation level, but that doesn't seem to help because the queries and updates involve more than one table.
I know that "reads don't block writes" is a fundamental part of the Oracle design that makes Oracle what it is, but is there any way I can explicitly make it happen anyway? Or can anyone see some other solution to what I'm trying to achieve?

Let me give a more real-world example to show what I'm talking about.
Imagine a simple bug-reporting OLTP application. Each bug has 3 fields:
1. Description
2. Resolution
3. Status (open/closed)
The application also has some rules:
A. If the Status is "closed", Description and Resolution cannot be changed.
B. In order to change the Status to "closed", both a Description and Resolution must be specified.
Now as long as only one person at a time updates a bug, everything is fine. However, if one person switches the Status to "closed" while another concurrently blanks out the Resolution you end up with a closed bug that has no Resolution. This is why some locking is necessary.
If Description, Resolution, and Status are all stored in the same table row, then it makes sense for each transaction to simply select the whole row FOR UPDATE, and then UPDATE it after it ensures all the rules are passed. However, in my fictional bug-reporting application they are all stored in separate tables. Now, the transaction that is making the change to Resolution could lock the row in the Status table FOR UPDATE (as everyone so far has suggested). Transactions making a similar change to the Description should then also lock the row in the Status table FOR UPDATE for the same reason. The end result of this is that Description changes and Resolution changes lock eachother out (can't happen concurrently, are serialized) when they don't need to be. Description changes and Resolution changes don't affect eachother, but because they both need to (exclusively) lock the same row, they are serialized.
If there was a way to acquire a share lock on the row in the Status table, then Description and Resolution changes would use that and not affect eachother. Status changes would still acquire an exclusive lock on the row, and would be blocked by (and block any further) concurrent Description and Resolution changes. This would be ideal.

SAP Lumira - Implementing row level security

Hi All,
I aware that SAP Lumira 1.17 onward allows to share the datasets, stories to SAP Lumira Server as well as SAP BI Platform (4.1 SP3 onward).
But I would like to know if there is any way of implementing Row level security for this published contents i.e. datasets or stories. e.g. If user A (may be an administrator with access to all the regions) creates dataset and story and shares it with other users over SAP Lumira Server or SAP BI Platform. But when user B accesses these contents on any platform, SAP Lumira server or SAP BI Platform, he should be able to see data only as per his access (his own region). Can something of this sort be implemented?
Thanks,
Abhijit

Hi,
Sorry for the delay in getting back to you.
As per my understanding - as of today, we respect Row-level security when acquiring (fetching) the data from universe into Lumira desktop (also, contexts and business-security profiles i.e. columns)
now, when that desktop user has 'designed' the Lumira document, all of the above: row-level, contexts and security profiles are 'locked-down' into that artefact when shared onwards. (i.e. to Lum Server and hence, BI Platform)
once this content is being access from the BI Launchpad, refresh-on-demand is possible from the story, as well as scheduling of dataset on which it is based.
According this blog by Greg Wcislo (the product owner for the Add-on) Lumira integration for BI4 functionality detailed. note that features such as 'refresh on open' and 'changing design-time parameters' (i.e. prompts) are not yet supported, but very much in future scope / plans.
I believe that one of the other mid-term goals is to architect a 'Lumira server-side universe refresh' (i.e. so that the processing is handled 100% by Lumira server) rather than querying across BIPlatform services then replicating a dataset to HANA (which is currently the process flow)
I hope this helps.
Regards,
H

Crystal Reports - ECC Tables - Row level security on Multiple tables

Hi Experts,
We are implementing Crystal Reports directly reporting on ECC Tables. Lot of information on row-level security has been provided by experts Ingo Hilgefort, Don Williamsand Mike Seblani, but not related to multiple tables or Wild cards
Requirement:
Crystal Users should have access to ALL the tables in ECC, but restricted by Company code, plant, Sales Organization, Purchasing Organization fields to what ever table it applies to. Example: MARC table should be restricted by Plant, BSEG table should be restricted by Plant and company code, GLT0 table should be restricted by Company code..etc
Users should ONLY see their Organization related data.
Solution Developed:
1. We created custom authorization object with BUKRS and WERKS
2. In /CRYSTAL/RLS we used Wild Cards *, + rather than specific table and referenced the custom authorization object with =BUKRS and =WERKS in the Field Value
3. Enabled global lock
4. Custom Authorization object was added to user-profiles with corresponding restrictions
*Observation:*
1. This security works when a crystal report was developed on a ECC table which has both BUKRS and WERKS
2. This setup DOES NOT work when a crystal report developed on a table with either one of BUKRS or WERKS
Example: Does not work on MARC table - error message "Database connection error: /CRYSTAL/OSQL_EXECUTE_QUERY Message: field T0~BUKRS" unknown"
Does not work on GLT0 table - error message "Database connection error: /CRYSTAL/OSQL_EXECUTE_QUERY Message: field T0~WERKS unknown"
Trouble Shooting:
In the "where clause" of the internal ABAP code generated for MARC, system is checking for BUKRS - which should not be the expected result
ANYTHING WRONG IN THE SECURITY SETUP ? PLEASE ADVICE
Note: Document "BusinessObjects XI Release 2, Integration Kit for SAP, Installation Guide" does not talk much about this multiple table restriction. Any other document to be referred to ?

I'm not sure how that would help; by using the Faculty_ID Session Variable I can identify the CRN and Term of all courses a faculty member is teaching. But I don't think that has to do with the problem I am having?

Remarks field in row level

Hi Experts,
My customer asked me the following question.
He uses the sales order document and write remarks in the free text field in the row level. When he is copying part of the quantity he will not able to change the free text field any more.
This field contain general remarks for the current row.
How can he cange row remarks after he copies part of the quantity.
The udf's in the row level got locked also.
Does anybody have an idea to solve it,
Thanks,
Joseph

Hi,
If I use copy from the fields will also get locked .
Is there any possibility to solve it. It's very important.
Thank you.

Row Level Security not working for SAP R/3

Hi Guys
We have an environment where the details are as mentioned below:
1. Crystal Reports are created using Open SQL driver to extract data from SAP R/3 using the SAP Integration Kit.
2. The SAP roles are imported in Business Objects CMC.
3. Crystal Reports are published on the Enterprise as well.
3. Authorization objects are created in SAP R/3 and added as required for the row level security as mentioned in the SAP Installation guide as well. The aim is when the user logs into the Infoview and refreshes the report he should only see data that he is meant to so through the authorization objects.The data security works very much fine when the reports are designed directly on the table but when the reports are built on the Business View it doesnt work hence the user is able to see all data.
Any help in this issue is greatly appreciated.
Thanks and Regards
Kamal

Hi,
In order for row level security to work for you using the OpenSql driver, you need to configure the Security Definition Editor on your SAP server. This is a server side tool which the Integration solution for SAP offers as a transport.
This tool defined which tables are to be restricted based on authorizations.
However since you are seeing the issue on reports based on Business Views, you need to identify whether the Business View is configured in such a way where the user refreshing the report is based on the user logging into Infoview. If the connection to your SAP server is always established with the same user when BV is used then you security definition is pointless.
You can confirm this by tracing your SAP server to identify what user is being used to logon to SAP to refresh the reports.
thanks
Mike

Row Level Restrictions on DeskI Report not working under InfoView

When running/refreshing DesktopIntelligence XI r3 (Full Client) reports under InfoView, the row-level restrictions applied in the universe are not being included in the SQL (additional filtering is NOT happening). However, if a user creates a report using the same universe in WebIntelligence or if the user runs the same report in DesktopIntelligence, the row-level restrictions are working.
It strikes me that the problem has to do with security rights granted (or not granted), but I can't figure out which settings would allow the proper running of a DeskI report under InfoView.
Thanks for any assistance rendered!

Mike,
Do people connect to Infoview and Deski using the same authentication method and the same username?
How does this compare to how your data restrictions are configured?
Are you using @Variable('BOUSER') restrictions? Have you taken domain into account e.g. "Domain\UserName" vs "UserName"?

Relationship between header and row level data.

Hi all
I need to know how to relate a field in the header level and the row level...
On my row level i want the route id to be related to the route name field at the header level...instead of showing this route id it must show only the header so...based on the row level data the header must change...
I tried insertin the field but when i do a relate to on either fields its showing me only fields on that catagory ( Header only or Row Level only)
Thanks

The Header is evaluated and printed first and then only Repetitive area is computed and printed. You cannot try to arrive header based on the Repetitive area.
Please explain in detail which screen etc.. so that we can suggest some work arounds

SBO - Approval Procedures at row level

Does anyone know how to create an approval procedure that is triggered by a row field. For example if I want an approval procedure that verifies discount percentage in row level.
There must be a workaround, cause I know this is not a functionality in SBO.
Thanks,
Hernán Baudrit

The way you could solve this is to create a UDF at Header level and link a formatted search which loops through the number of rows in the transaction and returns TRUE if for example the Discount % is greater than allowed.
The coding should be something as follows:
Declare @Counter as integer
Set @Couter = 0
Declare @TempTable TABLE (STAT varchar(50)) -- Temporary table holding results of loop
Declare @LoopFor as integer
Set @LoopFor = (select (*) from (select T1.DocEntry
                                                          from INV1 T0 join OINV T1 on T0.DocEntry = T1.DocEntry
                                                                  where T0.DocNum = $[OINV.DocNum])
While @Counter < @LoopFor
BEGIN
      Declare @Disc as integer
       Set @Disc = ( select T1.DiscPrcnt
                                        from INV1 T0 join OINV T1 on T0.DocEntry = T1.DocEntry
                                                   where T0.DocNum = $[OINV.DocNum]) and T1.LineNum = @Counter
     IF @Disc > 20
     BEGIN
              INSERT @TempTable VALUES ('TRUE')
     END ELSE INSERT @TempTable VALUES ('FALSE')
@Counter = @Counter + 1
END
SELECT Distinct 'TRUE' from @TempTable where STAT = 'TRUE'
Please note that the coding is not 100% accurate but should give you an idea of how to work around the problem.
Next you should write a simple validation query on the value of the UDF to check whether it is set to 'True'.
I hope this helps.

How to restrict table data at a row level?

Hi everybody,
I need create a database user and restrict his data access to certain tables.
What is the best way to do this?
Eg Table Emp has two columns, sal and dept. Sal has three distinct values: 500, 1000, 1500.
I need to make it so that new_user can only see the data in table emp where the sal value = 500.
Many thanks
Rup

You can use row-level security (VPD) for this sort of thing, so you assign a policy to the table that restricts which rows a particular user can see.
You can also roll your own solution by creating a view on the table and granting the user access to that view rather than direct access to the table.
Justin

Flex updates at a row level in a grid

I needs to updates row level in a grid for frequent basis, Also i don't want to Refresh the Grid.
Is there any method i can use ?
Using flex Grid -> updates at a row level in a grid

I mean DataGrid. I am trying to change the data rows based on realtime data feed.
First time i'll add all the Employees in the Grid. Later i'll get indivual request will change only Status column.
Is there is any other way i can update with out refresh in datagrid.
<mx:DataGrid
id="dg" height="260" width="900" x="0" y="20">
<mx:columns>
<mx:DataGridColumn headerText="EmpID" dataField="EmpID" width="10" visible="false"/>
<mx:DataGridColumn headerText="Emp Name" dataField="Emp Name" width="110"/>
<mx:DataGridColumn headerText="Status" dataField="Status" width="80"/>
</mx:columns>
</mx:DataGrid>

Row level locks

Similar Messages

Maybe you are looking for