Exit stored procedure after UPDATE

Question

Hello,
I have written a LOOP that calls a second Stored Procedure.
In the second Stored Procedure I have 3 INSERTS each SELECT's values limited with a WHERE clause.
Example:
begin
dynsql:=INSERT INTO destinationtable (ID,DATE,NAME)
SELECT src.ID,src.DATE, dest.NAME FROM sourcetable src INNER JOIN destinationtable dest ON dest.ID=src.ID WHERE dest.date between '01.JAN.2012 and '03.JAN.2012' --if not condition not met, no insert
execute immediate dynsql;
dynsql:=UPDATE destinationtable SET DATE = dest.DATE FROM sourcetable src INNER JOIN destinationtable dest ON dest.ID=src.ID WHERE dest.date between '01.JAN.2012 and '03.JAN.2012'
execute immediate dynsql;
dynsql:=INSERT INTO destinationtable (ID,DATE,NAME)
SELECT src.ID,src.DATE, dest.NAME FROM sourcetable src INNER JOIN destinationtable dest ON dest.ID=src.ID WHERE dest.date between '01.APR.2012 and '01.MAY.2012'  --if not condition not met, no insert
execute immediate dynsql
dynsql:=UPDATE destinationtable SET DATE = dest.DATE FROM sourcetable src INNER JOIN destinationtable dest ON dest.ID=src.ID WHERE dest.date between '01.APR.2012 and '01.MAY.2012'
execute immediate dynsql;
dynsql:=INSERT INTO destinationtable (ID,DATE,NAME)
SELECT src.ID,src.DATE, dest.NAME FROM sourcetable src INNER JOIN destinationtable dest ON dest.ID=src.ID WHERE dest.date between '01.JUN.2012 and '01.JUL.2012'  --if not condition not met, no insert
execute immediate dynsql;
dynsql:=UPDATE destinationtable SET DATE = dest.DATE FROM sourcetable src INNER JOIN destinationtable dest ON dest.ID=src.ID WHERE dest.date between  '01.JUN.2012 and '01.JUL.2012'
execute immediate dynsql;
end;
commit;What do I do?
I update the existing "old" row with a new date and then I am basically duplicating that row but adding it with a new DATE, the ID and NAME is "copied" as it is.
The big problem is that the rows result in recursive updates. if the first statement updates a date it might make the conditions for the second and first update become "true".
Which is not what I want. Initially, only one of the statements is true, ever. The problem with my code is that after the first update and date setting, another "might" become true.
This should stop. After one of the updates occured, I need to stop the stored procedure.
How do I stop it after upate?

Answer 1

metalray wrote:
Hello BlueShadow, that makes sense. Thanks.
Can you tell me when sql&rowcount is reset, after an execute immediate, after a commit?It doesn't get 'reset' in the strict sense. When a PL/SQL program unit is initiated the implicit "SQL" cursor space is defined and exists in the scope of that program unit. At that point it is like any variable that is declared with it's value undefined (null). When you issue any DML statement (insert, update, delete, merge etc.), these will explicitly set the ROWCOUNT attribute within the SQL cursor space. Other SQL statements are also issued from PL/SQL through the SQL "space" (interface), and as such, they may (it's not defined anywhere) effect the values of attributes within it, including the ROWCOUNT, but we cannot say exactly in what way. Therefore, the only thing that can be definitely determined is that a) when it's first instantiated, the SQL%ROWCOUNT will be null, and also, )_immediately_ after any appropriate DML statement the SQL%ROWCOUNT will contain the number of rows effected. If the SQL%ROWCOUNT value is examined at any other point, the value is indeterminate so, for example, after a commit statement, we cannot be sure what value it has or what that value refers to.
I loop through each tax no. need to do either an update, update or delete, delete, each with different WHERE clauses. the danger is that
I update and then the where clause for the deletion becomes true, but my process should not do the deletion if an update happened. its a sequential check and ones true and update/deletion made, the loop stored procedure should terminate. right now I check for sql&rowcount to "not" get into a update or delete if one already happened but I fail to grasp the logic of when sql&rowcount gets which value . I have the feeling that the parent stored procedure that calls my second one with the update and deletes sets the sqlrowcount to 1 because it has an insert. (I use 11g)After your insert/update you should immediately get the value of SQL%ROWCOUNT into a variable, so that you have the value before it is effected by other statements.
You can then use that value in your variable to determine program logic e.g. whether the delete should process or not, regardless of whether you issued further code between the insert/update and the check to see if you should delete. Don't rely on things 'resetting' the ROWCOUNT attribute, only rely on things setting it.