insert/update

Similar Questions

• SQL query generated by insertion/update of data vi?

  Hello!

  I'm new to LabView, now rated as my company's management plans to buy licenses for several developers. I'm now playing with the database connectivity toolkit and try to insert/updated updated data sets. Everything is good with simple tables, but when I try to execute relatively complex queries are often mistakes LabView diagnosed everything as "syntax error in Update statement" or something like that with no specific details.

  Is it possible to read the actual SQL queries generated by these vi or is there a (preferably free) tool to monitor all SQL queries called by my LabView program?

  Here is the link I promised last night. It includes a review of the issues with the CSD and why it is not necessary. There is also a link with some drivers that allow you to do what you need to do and are free - and a brief tutorial on the access to databases.

  Mike...

• Performance of insert/update SQLite

  Hi guys,.

  I tested module database in our application. I'm quite satisfied with the speed of select queries. I took ~0.1-0.3 seconds for large queries. But the insert/update quries, to 2-3 sec per operation. You have the same numbers?

  I read on temporary tables. Do you know other ways to speed up insertions/changes?

  Thank you

  Eugen

  I did a few tests.

  So packing several updates significantly increase the time of:

  20 insert/update statements ~ 45-53 seconds

  20 insert/update statements in a single transaction ~ 4-8 seconds

  This is an expected according to expirience android behavior.

  Also I tried TEMP tables really looks like fast ~ 1s. But I'm not sure about the test conditions "clear."  I'll play more with him.

• What is the best way to store the RCS for an insert/update in this rec

  Oracle on Win 64 non-conteneur 12.1.0.2

  When a record in one table is inserted or updated, what would be the best way to store the RCS for this record in this folder.

  I thought of a line after trigger, but did not know if this trigger to store the current_scn would still fire that trigger again (recursive trigger).

  Someone at - he a good idea of what the best way is to do?  The devs don't want to store the pk and the SNA in yet another table...

  Yes, row_dependencies would be the best way to go.  But mgmt doesn't recreate all tables for this.

  3rd party applications retrieve data from tables (all data).  We are looking for a way for them to just pull what is new or updated updated since their last sweater.

  I suggest that you try again and give all OF THE REQUIREMENTS.

  You have rejected ANY answer given and he justified using 'hidden' on what knowledge management or the devs want or do not want to. Stop making us guess what are the requirements and constraints. If you want a real answer then tell us ALL the news.

  When a record in one table is inserted or updated, what would be the best way to store the RCS for this record in this folder.

  Solomon answered repeatedly. If you want to add a column to a table to store the then "best" SNA is to let the Oracle to do this for you automatically by using the DEPENDENCY LINE.

  As he says also re-create the table to add this clause will be MUCH MORE EFFECTIVE that everything THAT you can do it manually. It will be also more accurate because Oracle will fill the value ORA_ROWSCN with the SNA at the time the line was committed. You, as long as user, can't fill a column in function when a line is engaged since real VALIDATION belongs to a transaction, not the line or the trigger that you use.

  Yes - there are two drawbacks to this method:

  1. you need to re-create the table

  2. you cannot add an index to this "hidden" column

  The devs don't want to store the pk and the SNA in yet another table...

  Then? Who cares what the devs want to do? You want the BEST solution? Next, you will need to put aside personal preferences and determine what is the 'best' solution. Why it is important that certain dev wants to do this or not?

  OK, the problem of biz is now, 3rd party external users are an all-wheel drive large number of tables in the database via the API that we wrote.  That was obviously interrupted OLTP during the day.  To reduce to the minimum, we want for them just to extract data that has been inserted/updated since their last sweater.

  It is the definition of a "replica" DB Then why don't you consider a real replicated DB? You can use DataGuard and have replicated DB which is read only that can be used to generate reports. Oracle does ALL the work to keep ALL the tables in sync. You and your developers do NOTHING!

  We thought that store the RCS higher their last sweater would allow the API to extract only data with YVERT higher than their last data pull CHN.

  OK - except you keep rejecting solutions actually do. Ask you questions about the SNA stored in the same table, but then reject the solution that does this. And then you add your "devs" don't want to store the info in a new table either.

  Then your solutions must ONLY use the replication or Log Miner. The REDO logs have all changes, if you want to extract yourself. Replication (e.g., DataGuard) will use these logs for you to maintain a replicated database.

  We thought about it, but recreate all tables in production with ROWDEPENDENCIES as well as dealing with CF and other dependencies idea this was shot.

  Well you NEVER mentioned you "thought that" and rejected it. And you NEVER mentioned anything about FKs and other dependencies. What is FKs and other dependencies which prevents this working solution? Tell us! Give us ALL the information.

  Wouldn't a trigger AFTER LINE capture the commit YVERT?  Or is after really not after validation?

  No - a trigger has NOT one commit. A trigger runs as a step in a transaction. Validation applies to the entire transaction. Until you, or Oracle, issues a commit, there is NO "committed SNA" to be stored as ORA_ROWSCN.

  You can easily see that for yourself. Create a simple table with dependencies of the line and then update two different sessions.

  create the table emp_scn rowdependencies in select * from emp where rownum<>

  Select empno, emp_scn ora_rowscn

  Update emp_scn set work = 'b' where empno = 7499

  commit;

  The first SELECT statement will show you that each row has the same SNA.

  EMPNO, ORA_ROWSCN

  7369,70622201

  7499,70622201

  7521,70622201

  Now, do the update (but no commit), then SELECT it

  EMPNO, ORA_ROWSCN

  7369,70622201

  7499,

  7521,70622201

  Where is the value of 7499? This session will NOT see a value for the changed lines in the current transaction. Other sessions will still see the old value.

  Now do the validation, then SELECT

  EMPNO, ORA_ROWSCN

  7369,70622201

  7499,70622301

  7521,70622201

  7499 now has a new and different value than the other lines. It will not be this new value until the validation occurs.

  Yes, row_dependencies would be the best way to go.  But mgmt doesn't recreate all tables for this.

  Well, you got the answer you want. You ask the best way. Now, you say that you were told the best way. But now you don't like the answer.

  How is it our fault? Your question has been answered wasn't she?

  Here are the facts:

  1 oracle creates a history of changes - the REDO log files

  2. you can use Log Miner to extract these changes

  3. you can create your own change log by adding a log file of MV to your table.

  4. you can then write a custom code to use this MV log file to determine which rows to "reproduce".

  So far reject you all THE POSSIBLE solutions.

  Accept it or change the requirements to allow one of the solutions proposed to be used.

  Personally, if I HAD to use a customized solution, I would use a MV journal to record the ROWID of the lines that have changed (for tables ROWID cannot be changed). I would then extract the appropriate lines by pulling on the lines corresponding to these row ID.

  Even that has problems since a line can be changed several times and children lines can also be amended several times - these questions FK you mentioned.

  I suggest you read this entire thread on AskTom a dozen years ago. It addresses ALL these issues.

  https://asktom.Oracle.com/pls/Apex/f?p=100:11:0:P11_QUESTION_ID:16998677475837

  Then in your next reply on this topic give us a summary of where some things with your question and what help you further expect.

• Check out when a last accessed table (select, insert, update) with time stamp. Auditing enabled

  Hello world

  IM pretty new to audit the database. Auditing is enabled in the database. I would like to retrieve the news all of the objects belonged to a certain pattern when it was last accessed (select, insert, update) with time stamp. Is there any script for this? Your time and your reply is greatly appreciated. Thank you.

  Database version: 11.2.0.4

  Enable audit is not quite enough to get the details when the table is updated/inserted/selected.

  You must activate audting on the object level, then you may only be able to see your report of your choice.

  SELECT OBJ_NAME, ACTION_NAME, to_char (timestamp, ' dd/mm/yyyy, hh') of sys.dba_audit_object.

• trigger check the previous record of line before insert/update

  Hi all

  I would like to create a trigger that a customer will not be able to create a new account, if it has a status of "Non-payment".

  CustomerID custNRIC paymentStatus

  ----------------------------------------------------------------

  Non-payment of 200 F7654323A

  It will reject the following statement.

  insert into customer (custID, custNRIC, paymentStatus) VALUES (201, 'F7654323A', 'Good');

  so, how can I go around to solve this issue?

  sqlnoob wrote:

  Hi all

  I would like to create a trigger that a customer won't be able to create a new account, if it has a "Non-payment" statusYou

  You can't do things like that in a trigger.

  It will reject the following statement.

  insert into customer (custID, custNRIC, paymentStatus) VALUES (201, 'F7654323A', 'Good');

  The trigger would dismiss as EVERY other row in the same transaction, if you try to use a trigger.

  You can't stop a trigger to run except by raising an exception and the trigger stop from doing anything for ANY line that the trigger activated on.

  together to define

  CREATE OR REPLACE TRIGGER reject_new_account

  AFTER INSERT OR update ON the client

  FOR EACH LINE

  DECLARE

  PRAGMA AUTONOMOUS_TRANSACTION;

  totover NUMBER (3);

  BEGIN

  SELECT COUNT (*)

  IN totover

  OF THE customer

  WHERE: NEW.nric = customer.nric

  AND: NEW.badstatus = "Non-payment";

  Why are you tring to use an autonomous transaction? All that makes the transaction will NOT be part of the transaction that fired the trigger.

  I know there is a problem with the line 13 ' AND: NEW.badstatus = "Non-payment".»

  so, how can I go around to solve this issue?

  This is the LEAST of your problems. Do NOT try to nontransactional things in a trigger.

  Also, don't try to query the table that the trigger is activated on.

  This is the fundamental question: Oracle is a multi-user system. Other users may be perform DML on the same table at the same time. So any NUMBER that provides this trigger can turn out to be inaccurate when other users post their transactions.

  The best you can do in a trigger is throw an exception. If you do this will affect the complete relaxation and all the lines he treated - not only one line you see that you want to reject.

  A requirement like yours needs to be satisfied by the presence of DML performed using functions or procedures and serialize the customer INSERT/update this transaction that a SINGLE session/user/can perform DML on the table for a client given.

• Unable to insert/update any registration via Oracle Forms 6i

  
  Hello

  I have a database block no (forms 6i) where I insert and update the application. Query is fired successfully, but it is not a document inserted or updated.
  Grant the user to insert/update.
  Commit is applied after each DML.

  I'm very surprised, why no data is not inserted into the database. Can anyone any suspicion that I might be missing?

  Other forms work well. All DML operations are working well in other forms, I get the error in a single form.

  There were a few exceptions at the level of the form. Then, form VALIDATION level did not work to validate the transaction DB, when I handled the exception, the validation of form level also worked to validate the DB transaction. Thanks to all for the helpful message.

• Insert/update the column with the clob data type

  Hi all

  ORCL Version: 11g.

  I have a table with the clob data type.

  Test12

  (col1 clob);

  I'm trying to insert/update to update the column with more than 4000 characters.

  But due to the limitation of tank 4000, I could not Insert/Update.

  Need your help in resolving this issue.

  THX

  Rod.

  The limit of 4000 characters is incorrect.  That pertains only to the varchar2 data type.  A clob can hold more than 4 G.

  Here is an example that shows how to insert it, I found...

  Otherwise, here is a way 'dirty' to do.

  insert into your_table (COLA, COLB)

  values

  (PRIMARY_KEY, PART 1 OF DATA)

  ;

  Update your_table

  Define COLB = COLB | PART 2 OF BIG DATA

  where COLA = PRIMARY_KEY;

  Update your_table

  Define COLB = COLB | PART 3 OF BIG DATA

  where COLA = PRIMARY_KEY;

  .. and so on...

  I don't know that I personally recommend the second style...  But he could do the job.

• insert/update - merge elements of apex

  Hello =)
  
  I want to insert data into a table, and if the table already has the "List_id' I want to update the values:
  
  Start
  BECAUSE me IN 1.APEX_APPLICATION. G_F01. COUNTING LOOP
  MERGE INTO list_values
  With the HELP of double
  WE (list_values.list_id = to_number (APEX_APPLICATION. G_F02 (i))
  WHEN MATCHED THEN
  UPDATE
  SET
  list_values.product_fk =: P900_PRODUCT_NAME.
  ist_values. Attribute = APEX_APPLICATION. G_F04 (i),
  list_values. Value = APEX_APPLICATION. G_F05 (i)
  WHEN NOT MATCHED THEN
  INSERT
  *(*
  list_values.product_fk, list_values. Attribute, list_values. Value
  *)*
  VALUES
  *(*
  *: P900_PRODUCT_NAME, APEX_APPLICATION. G_F04 (i), APEX_APPLICATION. G_F05 (i) *.
  *);*
  end loop;
  end;
  
  you have a lot of errors in sql...
  + 1 + error has occurred
  
  ORA - 06550:line 6, column 65: PL/SQL: ORA-00907: missing parenthesis ORA-06550: line 4, column 1: PL/SQL: statement ignored SQL
  
  Maybe someone can help me with the syntax of the merge statement or perhaps merging does not work in this case. because of the apex_items.
  
  
  also got separated insert/update process. they work perfect, if I use the process individually:
  Begin
  BECAUSE me IN 1.APEX_APPLICATION. G_F01. COUNTING LOOP
  
  insert into List_values (product_fk, attribute, value)
  values (: P900_PRODUCT_NAME, APEX_APPLICATION.) G_F04 (i), APEX_APPLICATION. G_F05 (i));
  END LOOP;
  end;
  
  Start
  BECAUSE me IN 1.APEX_APPLICATION. G_F01. COUNTING LOOP
  
  List_values update
  set
  product_fk =: P900_PRODUCT_NAME.
  attribute = APEX_APPLICATION. G_F04 (i),
  value = APEX_APPLICATION. G_F05 (i)
  where list_id = to_number (APEX_APPLICATION. G_F02 (i));
  END LOOP;
  end;
  
  Maybe someone can help me to combine.
  Thank you =)

  Name of the table on line 12 was also misspelled.

  BEGIN
     FOR i IN 1 .. APEX_APPLICATION.g_f01.COUNT
     LOOP
        MERGE INTO LIST_VALUES
             USING DUAL
                ON (LIST_VALUES.list_id =
                       TO_NUMBER (APEX_APPLICATION.g_f02 (i)))
        WHEN MATCHED
        THEN
           UPDATE SET
              LIST_VALUES.product_fk = :p900_product_name,
              LIST_VALUES.attribute = APEX_APPLICATION.g_f04 (i),
              LIST_VALUES.VALUE = APEX_APPLICATION.g_f05 (i)
        WHEN NOT MATCHED
        THEN
           INSERT     (LIST_VALUES.product_fk,
                       LIST_VALUES.attribute,
                       LIST_VALUES.VALUE)
               VALUES (
                         :p900_product_name,
                         APEX_APPLICATION.g_f04 (i),
                         APEX_APPLICATION.g_f05 (i));
     END LOOP;
  END;
  

  Jeff

• Is the sequence of the insert/update issue?

  Dear friends
  I wanted to know if there is a difference in the results, if I change the sequence of Insert/Update statements. For example, if I insert / update of the records in a table, normally I write my code like below
  
  < code >
  
  Begin
  Insert in the Table Name (...)
  Values (...);
  Exception
  When Dup_Val_On_Index then
  Update entire Table name
  Column_Name = Value;
  While others then
  SQLERRM;
  End;
  
  < code >
  
  Lets assume, if I change the order of statements like below:
  
  < code >
  
  Begin
  Update entire Table name
  Column_Name = Value;
  If Sql % NotFound then
  Begin
  Insert in the Table Name (...)
  Values (...);
  End;
  Exception
  While others then
  SQLERRM;
  End;
  
  < code >
  
  If this change makes no difference in the bottom line?

  I highly recommend the use of MERGE - make an update-so-insertion or insertion-so-update leaves you open to the clashes during two sessions try and add a new line at the same time.

  If you are not able to use MERGE for whatever reason, then I would work out who is most likely to happen - an insert or an update and do that first, followed by the other in the exception.

  For example. If your system inserts rows into the table in question a lot, but only updates rarely, so first the insert and manage the update in the exceptions. Conversely, if your system updates a lot but doesn't insert only rarely. In this way, you are minimizing the amount of work you have to do.

  Opt for a MERGER if possible, though. I don't know Oracle Forms, but can you not to call a stored procedure that could do the MERGE if you are not able to use MERGE directly in forms?

• Interface - value in the target field change on insert/update

  Hello
  
  For an interface, I use the module knowledge - incremental update of the IKM Oracle.
  
  He is running in insert / update of fashion.
  
  In the target of the interface table, there is a the target field (column) that I want to give a value based on exploitation.
  
  If the file is new and inserted into the target table - the field in the record should get the value "INSERT".
  
  -Similarly, if an existing record in the target table is updated - the field of the record should get the value "UPDATE".
  
  Any ideas on how you can do it?
  
  Thank you.

  You just need to customize your KM. Its can be done in many ways. Here is one approach.
  Change the existing lines of your update. Then change as below. But make you your own customization without hard coding.
  Here is codes, the fraction of the total code used in the KM steps.

  ){"?>(
       <%=odiRef.getColList("", "T.[COL_NAME]", ",\n\t", "", "((UPD AND (NOT UK) AND (NOT TRG)) AND REW)")%>
  *, COL_FLAG. *
  ) =
  (
  Select < %="odiRef.getColList" ("«,="" « s.="" [column] », »,="" \n\t\t\t »,="" « »,="" « ((upd="" et="" (pas="" uk)="" et="" (pas="" trg))="" et="" rew) »)="" %=""> *, "UPDATE."
  of< %="odiRef.getTable" ("l",="" "int_name",="" "w")="" %="" >="">
  where< %="odiRef.getColList" ("",="" "t.="" [column]="S." [column]",="" "\n\t\tand\t",="" "",="" "uk")="" %="">
  )

  Similarly for new lines inserted

  insert into <%=odiRef.getTable("L","TARG_NAME","A")%> T
  (
       <%=odiRef.getColList("", "[COL_NAME]", ",\n\t", "", "((INS AND (NOT TRG)) AND REW)")%>
       <%=odiRef.getColList(odiRef.getColList(",", " ", "", "", "((INS and !TRG) and REW)"), "[COL_NAME]", ",\n\t", "", "((INS AND TRG) AND REW)")%>
  *, COL_FLAG. *
  )
  Select< %="odiRef.getColList" (" »,="" « [column]="" «, »="" \n\t »,="" « »,="" « ((ins="" et="" (pas="" trg))="" et="" rew) »)="" %="">
  < %="odiRef.getColList" (odiref.getcollist="" («, »,="" "«,="" « »,="" « »,="" « ((ins="" and !="" trg)="" et="" rew) »),="" « [expression]="" «, »="" \n\t »,="" « »,="" « ((ins="" and="" trg)="" et="" rew) »)="" %="">*, "INSERT."

  of <%=odiRef.getTable("L","INT_NAME","W")%> S

  It should work. Your target must have the column as "COL_FLAG" and do nothing for this column in the map interface.
  Thank you

• Database insert/update by PLSQL having problems

  I have a piece of code that applies metadata for a set of file IDs. There are several metadata elements and multiple files (in general). So, basically, it's a loop to apply all the metadata is entered in a single file and then a loop to do the same to all the other files.
  
  The changes stem from a table of application that will contain information that is entered before you send the page by the user.
  
  The ID of the file comes from a colon delimited string, using a distribution function. If the string has been implemented 2456:8476:4939 str (2) would be 8476 and str (3) would be 4939. Function PLSQL itself has no significant errors, such as when compiling it works out, but when running it (in this case I as a process of page in the APEX) nothing happens, and when run outside of the application, he says one line was inserted each time, and nothing is inserted/updated in the database.
  
  I looked online and saw people have had similar problems, so I think it might be something simple that I'm just not.
  
  Here is the code:
  

  declare
      updatecheck VARCHAR2(2000);  --This will be used to check if an update or an insert is needed
      str string_fnc.t_array;                 --Array for the file IDs
      v_apex VARCHAR(2000);            
      countloop NUMBER;
  
  begin
  str := string_fnc.split(:P53_FILE_ID_LIST,':');                         --Splits the file ID string into an array.
  SELECT COUNT(METADATA_ID) INTO countloop FROM DD_METADATA_MASTER; --Determines how many pieces of metadata are currently possible to have
  for i in 1 .. str.count loop --For each file in the array, loop.
  for j in 1..countloop loop --For each piece of metadata that can be altered, loop.
  begin
  --If the current metadata item did not have new data entered, do not do anything for it.
  v_apex := APEX_APPLICATION.G_F01(j);
  exception
  when NO_DATA_FOUND then
  v_apex := null;
  
  end;
  --If it did have something entered, figure out if update is needed or not.
  if v_apex IS NOT NULL then
  begin
  --APEX_APPLICATION.G_F02(j) contains the ID of the metadata item. It is used to determine the type of metadata that is being used.
  --METADATA_VALUE_ID is simply a unique key whereas the master ID identifies which type of metadata the row is.
  
  select METADATA_VALUE_ID into updatecheck from DD_METADATA_VALUES where FILE_ID = str(i) and METADATA_MASTER_ID = APEX_APPLICATION.G_F02(j);
  exception
  when NO_DATA_FOUND then
  updatecheck := null;
  end;
  --If there is currently nothing in the values table for the current metadata item for the current file, new data should be inserted.
  if updatecheck IS NULL then 
        INSERT INTO DD_METADATA_VALUES (METADATA_MASTER_ID,METADATA_VALUE_ID,VALUE,FILE_ID) VALUES (APEX_APPLICATION.G_F02(j),DD_METADATA_VALUES_SEQ.NEXTVAL,v_apex,str(i));
  --If there already is metadata for the current item in the current file, update that data.
  else
  EXECUTE IMMEDIATE 'UPDATE DD_METADATA_VALUES SET VALUE = :1 WHERE FILE_ID = :2 and METADATA_MASTER_ID = :3' using v_apex,str(i),APEX_APPLICATION.G_F02(j);
  end if;
  end if;
  
  end loop;
  end loop;
  
  end;

  Published by: MichaelPaul on April 10, 2012 05:57

  MichaelPaul wrote:
  I tried a POSTING earlier, but let me test it in several places, rather than the only point that I had earlier to make sure that is not the issue.

  I tried with both, directly after insertion, and directly after the "EXECUTE IMMEDIATE"...

  Had similar problems, and that worked for me...

  HtH
  Johan

• Insert, Update, Delete on a partition table not even at a time

  Hello
  I have a non-partitioned table on which inserts, updates and deletes come at the same time due to which I am facing problem of contention.
  What would be the ideal solutions to solve my problem (deletion conflict.)
  
  Thank you
  Badin

  Can you give us more details on the "problem of contention"? This could mean a very large number of different things.

  You have different sessions simultaneously update the same row in the table, for example? If so, you probably need to look at the architecture of the application to avoid multiple sessions to the same line at the same time setting, or you would need to look at the data model to avoid multiple sessions to update the same line at the same time setting.

  But there could be dozens of other kinds of claim that you could talk.

  Justin

• Component ODI OWB 11 GR 2 How to connect lines inserted/updated to update / delete

  Hi experts,
  I developed an ICT KM which will insert into a target table data line-by-line, I used the ICT_SQL_INCR_UPD_ROW_BY_ROW KM of base provided by oracle, what I noticed, is that this newspaper does not work lines that are inserted as other KMs, after that a little research on the internet discovered that ODI has a function for this , oidRef.setNbInsert, but this doesn't work in OWB KM, it gives this error:
  --------------
  Traceback (innermost last):
  "< String >" file, line 305, inside?
  AttributeError: setNbInsert
  [12: 02:05] Szabo Adorian: org.python.core.Py.AttributeError (Py.java)
  org.python.core.PyInstance.invoke (PyInstance.java)
  org. $0 Python.pycode._pyx64.f (< string >: 305)
  org. Python.pycode._pyx64.call_function (< string >)
  org.python.core.PyTableCode.call (PyTableCode.java)
  org.python.core.PyCode.call (PyCode.java)
  org.python.core.Py.runCode (Py.java)
  org.python.core.Py.exec (Py.java)
  --------------
  
  Dose - anyone have an ideea how can I connect the inserted, updated, deleted in interface OWB lines (we have a customer who would like tho have this feature)?
  
  Enjoy a quick response,
  CIPI

  Hi CIPI

  Unfortunately, it is not included, the only possible audit for lines is via JDBC and properties explanations in KM.

  So not every API of ODI is supported, a large common subset is available, you can see the doc for more details (these audit would be useful);
  http://download.Oracle.com/docs/CD/E11882_01/OWB.112/e10935/code_tempates.htm#CIHIIJDI

  See you soon
  David

• With the help of Merge for Update, Insert, Update and also to delete outdated records

  Hi all!!!
  
  Suppose that a credit memo the following table called Orders:
  

  ORDERID        STATUS      PRICE 
  ---------- ---------- ---------- 
           1          0        100 
           2          0        200 
           3          0        300 
           4          0        350 
           5          0        390

  In addition, there are a procedure that will generate data on a global temporary table, called Orders_Temp. This temporary table will be used to update the Orders table. + commands + have a trigger that will perform different operations when an Insert, Update, or Delete command is executed.
  
  Consider the table of Orders_Temp to have the following data after you run a procedure that is responsible for the update of order data:
  

  ORDERID        STATUS      PRICE 
  ---------- ---------- ---------- 
           1          1        100 
           2          2        200 
           3          0        300 
           5          0        390 
           6          0        350

  As you can see, orders 1 and 2 have been updated. Order 3 and 5 were intact, order 4 has been deleted and command 6 inserted. Fusion will take care of properly manage orders 1, 2, 3, 5 and 6, although producing an update of order 3 and 5, which have not changed.
  
  So, here are the questions:
  (1) how could remove the 4 arrestedorder, as this order was not present on Orders_Temp? Is it possible to use the merge command to do this?
  (2) is it possible to generates not a update operation for orders of 3 and 5, which have not changed?
  
  I m using Oracle 10.2 SE.
  
  Thank you very much!
  Regis

  Hello

  Please try this. Not tested! with 9i. I think you can do this by using the union all.

  MERGE INTO orders tgt
   USING (SELECT A.*, 1 mask
            FROM orders_temp A
          UNION ALL
          SELECT A.*, 0
            FROM orders A
           WHERE NOT EXISTS
                   (SELECT 1
                      FROM orders_temp b
                     WHERE A.order_id = b.order_id)) src
      ON (tgt.order_id = src.order_id)
  WHEN MATCHED
  THEN
    UPDATE SET tgt.status = src.status, tgt.price = src.price
       WHERE src.mask = 1
         AND tgt.status != src.status
         AND tgt.price != src.price
    DELETE
       WHERE src.mask = 0
  WHEN NOT MATCHED
  THEN
    INSERT VALUES (src.order_id, src.status, src.price)
  

  or you could try to use a full outer join and treat NULL values.
  G.

  Edited by: g. March 8, 2011 11:17

  changed to the Union all the

• using postings in insert/update/delete

  Hello all;
  
  I'm so confused right now... I have read several articles on this subject and I don't know who is good or bad
  
  If an update statement is written in a client front-end as asp.net, must hire us or not, or should leave us it to auto-commit
  This scenario applies with an insert as well

  If you ask from the point of view of determining what is best practice

  Autocommit is evil and should be avoided at all costs. Unless you really, really, really understand what you do, a request that autocommits will end up with data in an inconsistent state. Unless the developers are extremely cautious, a page that allows you to transfer money from one account to another, for example, eventually sending statements to withdraw money from an account and drop it into the other and autocommit will cause to commit inadvertently withdrawal when filing statement fails.

  If you ask from the point of view of what are the technical requirements

  If you are in the mode autocommit, it is useless to engage in explicitly all the instructions (of course, you also won't be able to restore all the instructions explicitly)
  If you disable auto-commit mode, you will have to validate each transaction explicitly. After all never transactionally consistent of the DML statements (INSERT, UPDATE, and DELETE), you will need to do a commit explicit or an explicit rollback. In the example of transfer of account, you would hire after that withdrawal and the filing of declarations had been executed successfully. When you close a connection, your frame can do an implicit validation or cancellation (it depends on the frame) but who should not be trusted.

  Justin