ITextModelCmds::InsertCmd modifies the data before inserting it to the history

Similar Questions

• Check the data before inserting

  Hi all

  Appreciate your help. I have 3 tables, see below

  Examples of data

  CREATE TABLE PARENT_1

  (PARENT1_PK_COL NUMBER 4

  ) ;

  ALTER TABLE PARENT_1 ADD PRIMARY KEY

  (

  PARENT1_PK_COL

  )

  ;

  /

  CREATE TABLE PARENT_2

  (PARENT2_PK_COL NUMBER 4, NUMBER 4 PARENT2_FK_COL

  ) ;

  /

  ALTER TABLE PARENT_2 ADD PRIMARY KEY

  (

  PARENT2_PK_COL

  )

  ;

  /

  ALTER TABLE PARENT_2 ADD CONSTRAINT FOREIGN KEY ABC

  (

  PARENT2_FK_COL

  )

  REFERENCES PARENT_1

  (

  PARENT1_PK_COL

  )

  ;

  /

  CREATE TABLE CHILD_1

  (CHILD_1_PK NUMBER 4

  ) ;

  ALTER TABLE CHILD_1 ADD PRIMARY KEY

  (

  CHILD_1_PK

  )

  ;

  /

  My question is that before the insertion of data in the child_1 table, oracle look in parent_1.PARENT1_PK_COL and parent_2.PARENT2_PK_COL and columns if the data exists in both table then it allows to insert the record in the child_1 table, how do I apply this logic? do I need to add another foreign key in the child_1 table that references the two table? Please guide. Thank you

  Concerning

  MIT

  
  

  I got the answer, please ignore this thread

• Redirect data to another table using before insert trigger.

  Dear all,
  How can I redirect the data to another table in a before Insert trigger? My database is Oracle10g.
  I have a table EMP (EMP_ID, LAST_NAME, SALARY).
  I have another EMP_COPY table with the same structure. I also have a before Insert trigger on the EMP table.
  Based on a condition that I have to redirect the data in table EMP_COPY. Let's say the condition is EMP_ID = 100.
  I fire an insert on EMP table for example INSERT IN EMP(EMP_ID,LAST_NAME,SALARY) VALUES(100,'Dev',500).
  On the inside of the front Insert trigger on the EMP table, I have the code
  IF EMP_ID = 100 THEN
  INSERT INTO EMP_COPY (EMP_ID, LAST_NAME, SALARY)
  VALUES(:NEW.) EMP_ID,: NEW. LAST_NAME,: NEW. SALARY);
  COMMIT;
  ON THE OTHER
  NULL;
  END IF;
  But the problem here is that data goes to EMP table of origin also although I don't want. He should do an Insert into EMP table only if EMP_ID! = 100.
  
  One way has been to raise a user-defined exception inside the If statement and not handle it so that the original insert on table EMP fails but INSERT comes in the EMP_COPY table. But in this solution since the inside the trigger unhandled exception, it propagates to the calling environment. And I can't stand outside relaxation as the calling environment is a form of Oracle Apps standard that cannot be customized.
  
  Any kind of help will be highly appreciated that I am fighting for more than two weeks.
  
  Thanks in advance
  Dev

  Remove the autonomous transaction pragma... and then try again.

• Splitting words before inserting into the new table

  SQL> desc company
   Name                                      Null?    Type
   ----------------------------------------- -------- ----------------------------
   CO_ID                                     NOT NULL NUMBER(9)
   R_DATE                                             TIMESTAMP(0)
   CONAME                                             VARCHAR2(140)
   ESTYR                                              NUMBER(5)
   COUNTRY                                            VARCHAR2(32)
   PRODUCTS                                           VARCHAR2(3900)
  
  SQL> select unique(products) from company where rownum<6;
  
  PRODUCTS
  --------------------------------------------------------------------------------
  bars,tubes,pre-weld seams,isosceles and un-isosceles angles,special profiles,u-p
  rofiles,plates,t-profiles,flanges,tube fittings,round,hollow bars,double-t-profi
  les (ipe und ipb),pipe clamps,flat,special designs according to drawings for new
   products,profiles for floor covering,hexagon
  
  carbon steel bright bars,tool steel bright bars,alloy steel bright bars,stainles
  s steel bright bars
  
  cold rolled coil,e.g sheet,prepainted sheet,galvalume coil,copper scrap,stainles
  s steel sheet,shredded scrap,hms1 and 2,cobalt ore,rail scrap,hot rolled coil,ga
  lvanized sheet,iron ore
  
  PRODUCTS
  --------------------------------------------------------------------------------
  
  ferro alloy producer
  switch technology,special tubes,rails,steel strip,special steel

  I created another table:

  crate table keywords(
  k_id number(14),
  keywords varchar2(500) unique
  );
  
  create sequence keywords_pk_seq
  start with 3484 increment by 1;
  
  create or replace trigger keywords_pk_trig
  before insert on keywords
  referencing new as new
  for each row
  begin
  select keywords_pk_seq.nextval into :new.k_id
  from dual;
  end;
  /

  I want to:
  one) to split columns of company.products where is comma (,) for example, bars, tubes, pre-soudures, tubes will became 4 different as words:
  bars
  tubes
  welds
  tubes
  
  (b) then I want to insert these words in the table keywords.keywords with condition separated by ignoring the existing keywords.keywords (e.g. tubes came twice, let's get just a tubes) for example
  k_id keywords
  1 bars
  2 tubes
  3 before welding
  
  Thank you and best regards

  You can use the merge statement:

  for example

  merge into keywords
  using (
  with t as (select 'pipes,taps,spanner,spade,tubes' as txt from dual)
  -- end of sample data
    select distinct REGEXP_SUBSTR (txt, '[^,]+', 1, level) kw
    from t
    connect by level <= length(regexp_replace(txt,'[^,]*'))+1
    ) x
   on (keywords.keyword = x.kw)
  when not matched then insert (id, keyword) values (key_seq.nextval, x.kw)
  /
  

  But keep in mind that the sequence in this example will be updated for each line independently to know if she gets inserted or not.
  If you set the ID using a sequence triggered when inserting then this will solve this problem.

• Mask of the data before the select operation

  Hi all
  
  I have a requirement to hide data before the sound coming from the front. I have to do this only when users
  In addition to United Kingdom will access the system. Kindly let me know how I can do it.
  
  Thanks in advance.
  
  Best regards
  Pavan Poirier

  Hello

  Probably the simplest and the most reliable thing is to use a view instead of your actual table.

  How do you know if the current user is in the UK or not?
  If you have a function that returns "UK" when he passed the user name, then the view may look like this:

  CREATE OR REPLACE VIEW  view_x
  AS
  SELECT  *
  FROM    table_x
  WHERE   COALESCE ( user_country (USER)
                   , '?'
                   ) != 'UK'
  OR      ...
  ;
  

  , or, if the "mask" means to hide some columns as well as (or instead of) of rows, use the function in a CASE expression.

  In Oracle enterprise Edition, you can also use a security policy line level (also known as virtual private database or VPD) on the table (or view).

  I hope that answers your question.
  If this isn't the case, give more information. For example:
  "I have a table like this: CREATE TABLE...".
  with this data...
  When the United Kingdom users say SELECT * FROM... they should get these results...
  but the other users shou; to obtain these results...
  The way to know whether a user is on the UK's...
  My version of data is... »

• Loading data before the application starts

  Hello
  
  I have a complex application with lots of components and containers, some of whom require access to the data which I extract a DB via httpservice call.
  
  How can I make sure that the httpservice ended return data before the initialization of the application, so that the components that depend on data can be initialized correctly.
  
  
  Any help would be appreciated. Thank you

  Load your configuration data in the variables that may be related, bind your components of these vars (or use a ChangeWatcher).

  Tracy

• Truncate a field before inserting

  Suppose I have the following text:

  Array: mytable

  field: username (VARCHAR2 (35))

  field: zip (VARCHAR2 (5))

  I am importing data from an excel with columns username and zip file. A couple of the usernames contains more than 35 characters, and some postal codes are in the format xxxx-xxxxx, therefore containing more than 5 characters.

  I tried to write an insert before truncating the data until they are inserted. For any user name that is longer than 35 characters in Excel, simply truncate the username to the first of 35 characters. Similarly, for any ZIP code which is above 5 characters, just truncate the zip on the first 5 characters.

  The trigger for postal codes:

  CREATE OR REPLACE

  Trigger_zipc RELAXATION

  BEFORE INSERTING

  ON mytable

  FOR EACH LINE

  BEGIN

  SELECT substr (: new .zip, 1, 5)

  IN: new.zip

  Double;

  END;

  However, I get the error: ORA-12899: value too large for column

  I understand that this happens because the value is read before my validation, and because the value is too large, it is not allowed to insert.

  I'm not sure where to go from here. Any suggestions?

  Is it excel file or a csv file.

  I think you can look into creating an external table as well.

  Before inserting it just to make a substring.

  Something along the lines.

  CREATE TABLE ext_data (
   username CHAR(500),
    zip CHAR(500))
    ORGANIZATION EXTERNAL (TYPE ORACLE_LOADER DEFAULT DIRECTORY ext_tab_dir
    ACCESS PARAMETERS (FIELDS TERMINATED BY ',')
    LOCATION ('csv.dat'));
  
  insert into mytable
  select substr(username,1,35) , substr(zip,1,5) from ext_data;
  

  I hope this helps.

  Alvinder

• Before Insert TRIGGER to create partitions problem

  Hello
  
  I m having a problem with the following situation in Oracle 8i:
  
  I have a table TEST_TABLE, which is divided by the beach with a DATE column. The idea is to have a partition for each month, so the HIGH_VALUE of partitions is always the first day of the month following that represents the partition.
  
  I created a BEFORE TRIGGER INSERT on the table TEST_TABLE, which tests if the partition for the month of registration which is being inserted exists and, in case it doesn´t, a PROC AUTONOMOUS_TRANSACTION is called to create the TRIGGER.
  
  Running the code below one can see that even if partitions are created as expected, when you try to insert a record with a date greater than the last partition for the first time, this error is returned:
  ORA-14400: inserted partition key exceeds plu legal partition key.
  
  Note that if you run the same statement again insert, it s inserted correctly on the partition that was created the first try.
  
  I´ll appreciate any help on this matter.
  
  code
  ----------------
  CREATE TABLE TEST_TABLE)
  IDENTIFICATION NUMBER,
  DATE OF THE DT
  )
  TABLESPACE USERS
  PARTITION BY RANGE (DT)
  (
  PART_B42009 PARTITION VALUES LESS THAN (TO_DATE ('2009-01-01 00:00:00 ',' YYYY-MM-DD HH24:MI:SS ',' NLS_CALENDAR = GREGORIAN '))
  LOGGING
  TABLESPACE USERS
  );
  /
  
  CREATE OR REPLACE PROCEDURE SP_ADD_PARTITION (TEST_TABLE P_DATE. DT % TYPE)
  IS
  PRAGMA AUTONOMOUS_TRANSACTION;
  V_STR VARCHAR2 (500);
  BEGIN
  
  V_STR: = 'ALTER TABLE TEST_TABLE ADD.
  || 'PARTITION BIRD | TO_CHAR ("P_DATE, ' YYYYMM")
  || ' VALUES LESS (TO_DATE ("')).
  || TO_CHAR (ADD_MONTHS (P_DATE, 1), "YYYY-MM"). '-01 00:00:00 ','
  || ((' SYYYY-MM-DD HH24:MI:SS "," NLS_CALENDAR = GREGORIAN "))';
  
  EXECUTE IMMEDIATE (V_STR);
  
  END SP_ADD_PARTITION;
  /
  
  CREATE OR REPLACE TRIGGER TR_B_I_R_TEST_TABLE
  BEFORE INSERTING
  ON TEST_TABLE FOR EACH LINE
  DECLARE
  NUMBER OF V_PARTITION_EXISTS;
  BEGIN
  
  IF: NEW. DT > = TO_DATE ('2009-01-01 00:00:00 ',' ' YYYY-MM-DD HH24:MI:SS) THEN
  
  IMMEDIATELY EXECUTE (' SELECT COUNT (1) ")
  || "Of all_tab_partitions atp'."
  || "WHERE atp.table_name ="table_test"'.
  || "AND the atp. Nom_partition =: v1 ')
  IN V_PARTITION_EXISTS
  WITH THE HELP OF "BIRD" | TO_CHAR(:NEW.) "DT,"YYYYMM";)
  
  IF V_PARTITION_EXISTS = 0 THEN
  
  DBMS_OUTPUT. Put_line ('Partition [' |]) 'BIRD ' | TO_CHAR(:NEW.) "DT,"YYYYMM"). does not exist!') ;
  DBMS_OUTPUT. Put_line ('creation..');
  SP_ADD_PARTITION (: NEW.) DT);
  DBMS_OUTPUT. Put_line ('success.');
  
  IMMEDIATELY EXECUTE (' SELECT COUNT (1) ")
  || "Of all_tab_partitions atp'."
  || "WHERE atp.table_name ="table_test"'.
  || "AND the atp. Nom_partition =: v1 ')
  IN V_PARTITION_EXISTS
  WITH THE HELP OF "BIRD" | TO_CHAR(:NEW.) "DT,"YYYYMM";)
  
  IF V_PARTITION_EXISTS = 1 THEN
  DBMS_OUTPUT. Put_line ("it s visible at this point..");
  ON THE OTHER
  DBMS_OUTPUT. Put_line ("it s not visible at this point..");
  END IF;
  
  ON THE OTHER
  
  DBMS_OUTPUT. Put_line ('Partition [' |]) 'BIRD ' | TO_CHAR(:NEW.) DT, "YYYYMM")
  || already exists! ") ;
  
  END IF;
  
  END IF;
  
  DBMS_OUTPUT. Put_line ('continues with insertion...");
  
  END TR_B_I_R_TEST_TABLE;
  /
  
  -Go to the low score
  INSERT INTO TABLE_TEST VALUES (1, TO_DATE ('2008-12-31 23:59:59 ',' YYYY-MM-DD HH24:MI:SS'));))
  
  -Returns the error on the first try
  INSERT INTO TABLE_TEST VALUES (2, TO_DATE ('2009-01-01 00:00:01 ',' YYYY-MM-DD HH24:MI:SS'));))
  
  ----------------

  It is the use of the pragma AUTONOMOUS TRANSACTION. Your current transaction cannot see the result of this DOF since it occurs outside of the current transaction. The clue is in the name.

  Of course, you cannot run the DDL in a trigger without use of this pragma, so you're pretty much stuck. There is a solution in 11g, but that will not help you. Unfortunately, your only option is to pre-create the partitions required in front of the need. For example, you might have a DBMS JOB to create a partition for the next month, which takes place the last day of each month (or logical date of company).

  Cheers, APC

  blog: http://radiofreetooting.blogspot.com

• Clock set to date before January 1, 2001. Pop - ups require to accept incoming network connections for: krb5kdc, cups-lpd, MDNSResponder, configd. File system ARDAgent.app has been changed and cannot be repaired by disk utility. Files repaired permissions

  As soon as that connected with the password, the clock is set to date before January 1, 2001. Pop - ups require to accept incoming network connections for: krb5kdc, cups-lpd, MDNSResponder, configd. File system 'System/Library/CoreServices/RemoteManagement/ARDAgent.app/Contents/MacOS/ARDAg ent' has been changed and cannot be repaired by disk utility. Files permissions were repaired by disk on the install disc utility. How can I stop this?

  Am also using Firefox 43.0.3 browser like Safari.

  'System/Library/CoreServices/RemoteManagement/ARDAgent.app/Contents/MacOS/ARDAg ent' has been changed and cannot be repaired by disk utility. Files permissions were repaired by disk on the install disc utility. How can I stop this?

  You can ignore the above.

  Utility disk - permission errors, you can ignore

  System preferences/Date & Time/Date & time - is it set to automatically set the date and time?

• Historical research Google clears never when I clear the history browser and data

  I have a question that I barely noticed what is happening, it has never been a problem before. If I wanted to erase all of my internet history, (including searches on google), then all I do is go to settings, safari, clear history and data from the Web site, and it would be gone. Or even if I went to the history section on the safari browser whilse and clear of all time, he would go.

  However when I do it now, it still shows google searches!

  It is very annoying because I have to manually clear google searches, that clears only 3 search at a time.

  This AP is an ios issue, because it's never happened before.

  Your Google search history is stored on Google's servers. You should be able to erase by going to www.google.com (this on a computer), login and go to your search history. Access your account, your personal information, the search parameters. Down at the bottom of the page, you will see a "search story" a hyperlink. You can click on that and change your story here.

• Data before release for NOR-6133 trigger material and acquisition continue?

  I use the pxi of NOR-6133 can acquire data on 7 channels continuously with ai0 as the trigger without problems. The acquisition is triggered. What I have to do is collect a finite amount of data before triggering immediately until the trigger is detected.

  Is it possible to do without the help of a software solution. I watched go over acquisition and using the integrated trigger reference but that is not suitable for my needs because it just seems to change the first section of data to be 'post trigger' to 'pre trigger.

  A software solution that I already know how to do is the last resort to try to reduce the demand for processor and keep the hardware triggering.

  Thank you

  By using the "reference trigger" is the right solution.

  So can you please explain a little more in detail why it does not work in your case?

  Christian

• Confirm delete before inserting then have a unique key.

  Hello

  I have a use case where a database field that is entered by the user must be unique.

  In the database, there is already a record for example 'name1 '.

  The user inserts a new record and puts his name 'name1' as well. Then, for some reason any decides to delete the first existing record and tries to save the new record 'name1 '.

  The previous record validation "Name1" is not yet removed, but the new record with 'name1' seeks to insert.

  Then I get the error message: attribute value 'name1' leads to duplicate keys for this type of entity. oracle.jbo.AttrValException...

  Is it possible to run delete before insertion on commit or another work around?

  JDev Version 11.1.2.3.0

  Concerning

  Here is the link which suites for your usecase

  ADF binary Decompilation: customization of the display order of the changes made to the database

  You need to wtite your own custom of the Transactionfactory class.

  Let us know if it works for you

  Thank you

  KT

• TRIGGER BEFORE INSERT OR UPDATE OR DELETE

  I want to create a single trigger before each insert or update or delete, a record gets copied into the AUDIT table. works very well on the insertion and to day but you have problem clear.

  Can you please help me debug this error. The problem of mutation.

  create or replace

  BIU_FS_Test_T RELAXATION

  BEFORE INSERT OR UPDATE OR DELETE

  ON FS_Test

  REFERRING AGAIN AS NINE OLD AND OLD

  FOR EACH LINE

  DECLARE

  number of voffender_history_id;

  Start

  IF the INSERTION

  SELECT FS_Test_history_seq. nextval in the double voffender_history_id;

  INSERT INTO FS_Test_history (OFFENDER_HISTORY_ID,

  OFFENDER_ID, FIRST_NAME, MIDDLE_NAME, LAST_NAME.

  SX, SKIN_TONE, HEIGHT, WEIGHT, RACE, EYE_COLOR,

  HAIR_COLOR, BLOOD_TYPE, VARIOUS, DATE_OF_BIRTH.

  PLACE_OF_BIRTH, TRANSACTION)

  VALUES (voffender_history_id,

  : new. OFFENDER_ID,: new. First name: new. MIDDLE_NAME,: new. LAST_NAME,

  : new. RACE: new. SX,: new. SKIN_TONE,: new. HEIGHT: new. WEIGHT: new. EYE_COLOR,

  : new. HAIR_COLOR,: new. BLOOD_TYPE,: new. VARIOUS: new. DATE_OF_BIRTH,

  :new. PLACE_OF_BIRTH,

  "INSERT");

  ELSIF UPDATE

  et (

  NVL (:OLD. First name, 'x') <> nvl():NEW. First name, 'x') or

  NVL (:OLD. MIDDLE_NAME, 'x') <> nvl():NEW. MIDDLE_NAME, 'x') or

  NVL (:OLD. Last_name, 'x') <> nvl():NEW. Last_name, 'x') or

  NVL (:OLD. RACE 0) <> nvl():NEW. RACE 0) or

  NVL (:OLD. SX 0) nvl(:NEW <>.) SX 0) or

  NVL (:OLD. SKIN_TONE 0) <> nvl():NEW. SKIN_TONE 0) or

  NVL (:OLD. HEIGHT 0) <> nvl():NEW. HEIGHT 0) or

  NVL (:OLD. WEIGHT 0) <> nvl():NEW. WEIGHT 0) or

  NVL (:OLD. EYE_COLOR 0) <> nvl():NEW. EYE_COLOR 0) or

  NVL (:OLD. HAIR_COLOR 0) <> nvl():NEW. HAIR_COLOR 0) or

  NVL (:OLD. BLOOD_TYPE 0) <> nvl():NEW. BLOOD_TYPE 0) or

  NVL (:OLD. VARIOUS, ' x') <> nvl():NEW. VARIOUS, ' x') or

  NVL (:OLD. DATE_OF_BIRTH, January 1, 1900 ") <> nvl():NEW. " DATE_OF_BIRTH, January 1, 1900 ") or"

  NVL (:OLD. PLACE_OF_BIRTH 0) <> nvl():NEW. PLACE_OF_BIRTH 0)

  )

  THEN

  SELECT FS_Test_history_seq. nextval in the double voffender_history_id;

  INSERT INTO FS_Test_history (OFFENDER_HISTORY_ID,

  OFFENDER_ID, FIRST_NAME, MIDDLE_NAME, LAST_NAME.

  SX, SKIN_TONE, HEIGHT, WEIGHT, RACE, EYE_COLOR,

  HAIR_COLOR, BLOOD_TYPE, VARIOUS, DATE_OF_BIRTH.

  PLACE_OF_BIRTH, TRANSACTION)

  VALUES (voffender_history_id,

  : new. OFFENDER_ID,: new. First name: new. MIDDLE_NAME,: new. LAST_NAME,

  : new. RACE: new. SX,: new. SKIN_TONE,: new. HEIGHT: new. WEIGHT: new. EYE_COLOR,

  : new. HAIR_COLOR,: new. BLOOD_TYPE,: new. VARIOUS: new. DATE_OF_BIRTH,

  :new. PLACE_OF_BIRTH, "UPDATE");

  ELSIF DELETION THEN

  SELECT FS_Test_history_seq. nextval in the double voffender_history_id;

  SELECT FS_Test_history_seq. nextval in the double voffender_history_id;

  INSERT INTO FS_Test_history (OFFENDER_HISTORY_ID,

  OFFENDER_ID, FIRST_NAME, MIDDLE_NAME, LAST_NAME.

  SX, SKIN_TONE, HEIGHT, WEIGHT, RACE, EYE_COLOR,

  HAIR_COLOR, BLOOD_TYPE, VARIOUS, DATE_OF_BIRTH.

  PLACE_OF_BIRTH, TRANSACTION)

  VALUES (voffender_history_id,

  : new. OFFENDER_ID,: new. First name: new. MIDDLE_NAME,: new. LAST_NAME,

  : new. RACE: new. SX,: new. SKIN_TONE,: new. HEIGHT: new. WEIGHT: new. EYE_COLOR,

  : new. HAIR_COLOR,: new. BLOOD_TYPE,: new. VARIOUS: new. DATE_OF_BIRTH,

  :new. PLACE_OF_BIRTH, 'CLEAR');

  END IF;

  end;

  Error I get

  Error from the 1 in the command line:

  Remove from the fs_test where PK = 41476

  Error report:

  SQL error: ORA-04091: table get OUT. SOR_O_HISTORY is changing, function of triggering/can not see

  ORA-06512: at «GET. » BIU_FS_TEST_T', line 111

  ORA-04088: error during execution of trigger ' EB. BIU_FS_TEST_T'

  04091 00000 - «table %s. » "%s is changing, function of triggering/can not see.

  * Cause: A trigger (or plsql user-defined function that is referenced in)

  ( this statement) attempted to watch (or modify) a table that has been

  in the middle to be modified by the statement that shot.

  * Action: Rewrite the trigger (or function) so it does not read this table.

  PS 11g r2

  Thank you.

  n_shah18-

  Is there a parent-child relationship between the tables involved in the triggering event? Is there a foreign key used above the child table that references the parent table?

  If you have a parent-child relationship and a trigger on the parent table must be running some DML on the child table, and, if the child table as a foreign key (FK) back to the parent table, any DML on the child table will cause a recursive SQL statement to check the constraint. Indirectly, this will cause an exception table mutation.

  So watch the dependencies on SOR_O_HISTORY / FS_TEST_HISTORY/all other related table. A FOREIGN KEY can make it to mutate (no pun intended).

  Hope this will help you find the root cause.

• Conversion of date and insert

  Hello
  We have a small script that we need help please.
  
  We treat the txt files in a plsql code, the name of the file is usually in the format "SDE_NN_YYYYMMDDHHMI.txt".
  
  The NN is a 2 digit number
  The YYYYMMDD is the date of the file
  HHDI is the component of hour and minutes.
  
  While we treat the data found in the txt file, we must update the underlying table with the date in the name of the file.
  
  We use the current code to do:
  

  FILE_DATE := to_date(substr('SDE_20_201205071235.txt',instr('SDE_20_201205071235.txt','_',1,2)+1,8), 'yyyymmdd');

  The code works well, and it does not extract the date component out of the string of file name without any problems. The issue that we have is that even if the date is to be extracte with success, once inserted, then ask the changes are committed, if we come back and do a simple select on the field of file_date, we get the current sysdate instead of the actual date of inserted (based on the name of the file)?
  
  We even tried the following syntax for the insertion of the value date:
  

  insert into contributions values
  (con_seq_num, SDE_TYPE_ID, SDE_ID, oDEPOSIT_DATE, SDE_AMOUNT, AGGREGATE_AMOUNT, SDE_YEAR, NAME, sysdate, SDE_NAME, FILE_NAME, to_date(substr(u_in_fname,instr(u_in_fname,'_',1,2)+1,8), 'yyyymmdd'), null, null, 'sde_ADMIN',  null); 

  But once more, the insert is committed successfully, and when we run the select, we get the current sysdate instead of the desired date (which in this case should be may 7, 2012 (20120507)).
  
  Any ideas please?
  
  Version Oracle 11.2.0.1
  O.s Windows 2008 server.
  
  
  Thank you

  There may be a trigger on the table that is inserting SYSDATE.

• Trigger before insert

  Hi, I need help to create a trigger before the insertion of the STA_STARTDATE column in the employees table
  

  CREATE OR REPLACE TRIGGER trig_start_date
  BEFORE INSERT OF STAFF FOR EACH ROW
  BEGIN
    IF STA_STARTDATE < '01-JAN-1980'
      THEN
          RAISE_APPLICATION_ERROR(-20900, 'No earlier than 1/1/1980');
    END IF;
  END;

  Note: If the declaration is < 01 - JAN-1980 quotes
  
  When I run the script, it comes up with this error:
  
  Error report:
  ORA-04073: invalid column list for this type of trigger
  04073 00000 - "list of invalid columns for this type of trigger.
  * Cause: A list of columns has been specified for a non updated trigger type.
  * Action: Delete the column list.
  
  * fixed by replacing the OF keyword with, but still have a warning
  
  Published by: 888948 on October 2, 2011 16:48

  Hello

  to use a column in a trigger, the column must have: new for new values, or: old old values before the name of the column, so your if statement should begin with:

  IF :NEW.STA_STARTDATE 
  

  Herald tiomela
  http://htendam.WordPress.com