Exception Handling of Screenflows inside Processes

Greetings,
We are currently encountering a problem where some unhandled Exceptions are not being caught in our BPM application.
Whenever an exception occurs inside a process, BPM has provided a way of handling those exceptions by using the “Process Exception” activity (this looks like a small lightning bolt). All errors that are not caught but are of the compatible type would fall into this activity and get processed.
The problem arises when your process has a Global Activity Interactive Screenflow. Whenever this screenflow encounters an error inside its flow, it does not get caught by the “Process Exception” activity even though this process was the one that invoked it.
We can try to make some exception handling per automatic activity inside the Screenflow itself but this usually results in large try-catch statements and multiple connections to custom defined exception activities. This may not be the optimal solution to this.
Would anyone be able to shed some light on this or offer some best practices on how to resolve these issues?
Thank you very much!
Oracle BPM Suite 10.3.2

Wish I had a better answer for you. Telling you something that you already know, but if you've defined the Global's property with "Has instance access" then you can catch the exceptions in the process.
Undoubtedly though, you've got a standalone Global or you would not be asking the question. The exception from a Global that does not have instance access can't get caught in the process because there was never an instance created for it inside the process.

Similar Messages

Exception Handling inside a Multi-Instance Loop

I would like to see a sample process that demonstrates Exception Handling inside an inline subprocess containing another subprocess with a multi-instance sub-process in parallel. The outer sub-process is executing in sequence. Each instance of the outer loop depends upon the outcome of the successful execution of previous step. At each step, the inner inline sub-process activity can have more than one instance which are all executed in parallel using multi-instance. If the outcome code of any one of these parallel instances is "REJECT" code, we simply raise a business exception and the stop the outer sub-process from going through the next instance of the loop. The problem we are trying to solve is similar to the sample in chapter 5 of the book "New Book: Oracle BPM Suite 11g: Advanced BPMN Topics" by Mark Nelson. Particularly, the exception handling example shown in Page 73 under the topic “Exception Handling with embedded Sub-processes”. The inner most multi-instance sub-process should raise a business exception and interrupt the .
We would like to see a smple that demonstrates how exceptions are handled inside a multi-instance parallel sub-process. Could someone please provide a working sample that we can go though? We would like to raise a business exception as soon as certain outcome of a Human Tasks in observed and break out of the loop and continue thereafter. Thanks very much in advance for your help.
Pankaj
Edited by: 1001027 on 2-May-2013 10:09 AM

I would like to see a sample process that demonstrates Exception Handling inside an inline subprocess containing another subprocess with a multi-instance sub-process in parallel. The outer sub-process is executing in sequence. Each instance of the outer loop depends upon the outcome of the successful execution of previous step. At each step, the inner inline sub-process activity can have more than one instance which are all executed in parallel using multi-instance. If the outcome code of any one of these parallel instances is "REJECT" code, we simply raise a business exception and the stop the outer sub-process from going through the next instance of the loop. The problem we are trying to solve is similar to the sample in chapter 5 of the book "New Book: Oracle BPM Suite 11g: Advanced BPMN Topics" by Mark Nelson. Particularly, the exception handling example shown in Page 73 under the topic “Exception Handling with embedded Sub-processes”. The inner most multi-instance sub-process should raise a business exception and interrupt the .
We would like to see a smple that demonstrates how exceptions are handled inside a multi-instance parallel sub-process. Could someone please provide a working sample that we can go though? We would like to raise a business exception as soon as certain outcome of a Human Tasks in observed and break out of the loop and continue thereafter. Thanks very much in advance for your help.
Pankaj
Edited by: 1001027 on 2-May-2013 10:09 AM

Integration Process exception handling & alerts, Java

Hello,
I would like to enhance my basic scenarios fool proof and with as much Java as possible.
I have set up several asynchronous and synchronous File to SOAP transfers between systems A and B
using PI70 including my own Java mapping classes and PI's AF_Modules beans. They work just fine.
I have not designed my own Integration Processes yet, so execution is based on channel settings and availability timing.
I have learned this aproach may be exposed to general faults resulting manual monitoring and repairing.
For example server problems at receiver side can result data losses as sender channel just keeps removing source
files like normally, as the process would be better to just stop right there and alert or something.
My question is, is the "Enterprise services Builders" Integration Process & Graphical definition screen the one and only tool to
customize whole process exception handling and alerts in PI, and Java is not an option?
Appreciate your advice on this.
Kind regards m

Hi m,
Strange Name
>>I have learned this aproach may be exposed to general faults resulting manual monitoring and repairing. For example server problems at receiver side can result data losses as sender channel just keeps removing source files like normally, as the process would be better to just stop right there and alert or something.
If you are looking to handle this particular scenario, then we have the alert mechanism and CCMS monitoring. There you will come to know whether the end system down or not.
Also in addition to this you can write your own java (in message mapping, as modules, java mapping) /abap code (as abap mapping, user exit in standard functions etc) for providing more details in error scenarios. But you need to validate whether the maintenance/development cost for the code is justified in your scenarios
Regards
Suraj

BPM Process - Exception handling or timeout issues?

Hi Guys,
I have a BPM process as below.
1. Receive step: Receive the file with multiple transactions.
2. Transformation step: Split the file into individual transactions
3 Block step which includes -- par for each mode
1. Send Step (Synchronus): Each individual transaction needs to contact the 3rd party system and get the response. -- Do i need to handle any exceptions here ?
2. Container : Collect all the responses
Block ends
4. Transformation: combine all the responses in to a single file
5. Send Step: synchronus -- send the above single file and get the response back
6. Transformation : Transform the above response into the target structure.
7. Send: send the message asynchronusly to the target system
I need suggestion regarding the exceptional handling or any time out issues, i need to take care of.
any suggestions would be really appreciated
Thanks,
Raj
Edited by: raj reddy on Feb 12, 2009 10:12 PM

Hi,
I) For the Block holding the Sync Send, create an Exception Block. (right click on Sync Send -> Insert -> Exception Branch)
II) Name the Exception block (ex: exceptionHandler).
III) in the Sync Send step ->Properties -> Exceptions -> in System Error - add exceptionHandler.
IV) Now within the Exception handler block you can create containers to hold values from payload, throw exception as email etc).
This will cover your sync send step incase there is an error while sending the request of a timeout during receiving the response.
You can also do the same for the Step 7) Asycn send - if required.
Another suggestion in your question Step 6) can be done outside the bpm, when you do the interface determination for that Asycn Send you can add the Interface mapping that will map the responses to the target structure.
Doing this will reduce one step in your BPM. For further information in how more you can fine tune your bpm, read this blog - https://www.sdn.sap.com/irj/scn/weblogs?blog=/pub/wlg/5113. [original link is broken] [original link is broken] [original link is broken]
All the best.
Regards,
Balaji.M

Exception handling in integration processes

Hi,
I have a very simple question but I can't find an answer. I have designed an integration process that upon several other things send synchroneously a message to an ABAP proxy. Sometime the ABAP proxy generates exception. In my integration process these exceptionx are catched in a specific branch.
My question is: how to access the data contained in the catched exception (message, text....)?
Regards,
Yann Le Blevec

Hello Yann,
There is no solution for your problem, I guess. I didnt try it on my own, but in my own research on sdn I found two experts positions to that issue:
"you cannot catch the exception's value in the XI "
(see posting from Michal Krawczyk, Jun 12, 2006, in thread catch the exception)
"The content of a fault message are not accessible inside a BPM."
(see posting from Bhavesh Kantilal, Oct 3, 2007, in thread Notify user about Exception in BPM)
Regards,
Dennis

Resume process instance after exception handling with an event subprocess

We have a process with several automatic activities so we implement an event-based subprocess to catch any exception that occurs in the process.
Since the error event is an interrupting event, how could we control the exception and resume the process where it was?
BPM version 11.1.1.7
Any suggestion?
Thank you

Know this isn't the answer you are looking for, but the answer to this gets much better in 12c where you can go back into the process after catching an exception in an event subprocess.
In 11g however, once you catch an exception in an event subprocess you cannot go back where you left off in the main process.
These are not great options I know, but here are a couple things I've seen to work around this:
Exception error events in event subprocesses are always interrupting so one solution would be to instead use boundary events on the individual activities where the exceptions will occur. Know you know this, but the down side of this is that it makes your process incredibly cluttered with boundary events.
Catch the error in the event subprocess -> handle the cause of the exception -> exit the process -> invoke the process again. The problem with this is that you would not be starting where you left off and where the error occurred. You could work around this by having an exclusive gateway that directed it to the right activity in the process.
Dan

Exception handling in stored process, loop IF..ELSE

Hello Guys,
we want to put in exception handling in the loop but get the following error:
Error(43,3): PLS-00103: Encountered the symbol "EXCEPTION" when expecting one of the following: begin case declare end exit for goto if loop mod null pragma raise return select update while with <an identifier> <a double-quoted delimited-identifier> <a bind variable> << close current delete fetch lock insert open rollback savepoint set sql execute commit forall merge pipe
create or replace
PROCEDURE xxxx
FOR MESSSY IN
select I.*
FROM x I
LOOP
IF upper(CODE)='N' THEN
INSERT INTO T_MESS(MP)
select I.MP_ID
FROM T_ME
ELSIF upper(MESSSY.k2)='L' THEN
DELETE T_MESS WHERE T_MESS.MP = MESSSY.MP;
END IF;
EXCEPTION
WHEN DUP_VAL_ON_INDEX THEN
A program attempted to insert duplicate values in a column that is constrained by a unique index.
DBMS_OUTPUT.PUT_LINE ('A program attempted to insert duplicate values in a column that is constrained by a unique index.')
--No Rollback
END;
COMMIT;
END LOOP;
END xxxx;
does someone know why?

BluShadow wrote:
Well, your code is missing all sorts of bits and we don't have your data or your exact logic to know what it's supposed to be achieving.
That is right, you dont have my data and that is why I was suprised by your comment.
Since the input table might contain a few thousand rows and each of those might need to
be considered N , D, or C and each case has a different handling I can not imagine how this
can be all done with a merge statement.
MERGE
T_METRICPOINT_META with T_METRICSSYSTEM_LOAD where T_METRICSSYSTEM_LOAD .LOAD_DATE=to_char(sysdate)
WHEN MATCHED THEN --we know those are the metric points that have to be loaded today, but we still need to do a IF..ELSE to handle them
WHEN NOT MATCHED THEN -- not considered in todays load
----original code-----
create or replace
PROCEDURE myprocedure AS
BEGIN
--Extracting the records from T_METRICSSYSTEM_LOAD which have todays load date. Corresponding to these MP_System, we extract the MP_IDs from the T_METRICPOINT_META table.
--Comapring these MP_IDs with the MP_IDs from the source(T_METRICPOINT_IMPORT) and extracting only those Metric points which need to be loaded today.
FOR METRICSSYSTEM IN
select I.*
FROM T_METRICPOINT_IMPORT I
where I.LOADDATE = TO_CHAR(SYSDATE) AND I.MP_ID IN
(select a.MP_ID
from T_METRICPOINT_META a INNER JOIN T_METRICSSYSTEM_LOAD b on a.MP_SYSTEM = b.MP_SYSTEM where b.LOAD_DATE=to_char(sysdate))
LOOP
--If mutation code in the source/import data is "N", the record is inserted as it is in the "T_METRICPOINTS" table.
IF upper(METRICSSYSTEM.MUTATIONCODE)='N' THEN --new
INSERT INTO T_METRICPOINTS(MP_ID, ......)
SELECT DISTINCT I.MP_ID,.....
FROM T_METRICPOINT_IMPORT I WHERE I.MP_ID = METRICSSYSTEM.MP_ID
ELSIF upper(METRICSSYSTEM.MUTATIONCODE)='D' THEN --delete
DELETE T_METRICPOINTS WHERE T_METRICPOINTS.MP_ID = METRICSSYSTEM.MP_ID AND T_METRICPOINTS.KEY = METRICSSYSTEM.KEY;
ELSIF upper(METRICSSYSTEM.MUTATIONCODE)='C' THEN --correction
UPDATE T_HISTORYMETRICPOINTS H
SET CHANGE_DATE = to_char(sysdate)
WHERE H.MP_ID=METRICSSYSTEM.MP_ID AND H.KEY = METRICSSYSTEM.KEY;
INSERT INTO T_HISTORYMETRICPOINTS(MP_ID, KEY, .....)
--The distinct here is used, to handle 2 identical records in the input table with correction value "C". This would insert into 1 record in the T_HISTORYMETRICPOINTS table without
--violating the primary key constraint.
select DISTINCT I.MP_ID,I.KEY, ....
FROM T_METRICPOINT_IMPORT I WHERE I.MP_ID = METRICSSYSTEM.MP_ID
--END IF;
END IF;
COMMIT;
END LOOP;
END myprocedure;

PL/SQL 101 : Exception Handling

Frequently I see questions and issues around the use of Exception/Error Handling in PL/SQL. More often than not the issue comes from the questioners misunderstanding about how PL/SQL is constructed and executed, so I thought I'd write a small article covering the key concepts to give a clear picture of how it all hangs together. (Note: the examples are just showing examples of the exception handling structure, and should not be taken as truly valid code for ways of handling things)
Exception Handling
Contents
1. Understanding Execution Blocks (part 1)
2. Execution of the Execution Block
3. Exceptions
4. Understanding Execution Blocks (part 2)
5. How to continue exection of statements after an exception
6. User defined exceptions
7. Line number of exception
8. Exceptions within code within the exception block
1. Understanding Execution Blocks (part 1)
The first thing that one needs to understand is almost taking us back to the basics of PL/SQL... how a PL/SQL execution block is constructed.
Essentially an execution block is made of 3 sections...
+---------------------------+
|    Declaration Section    |
+---------------------------+
|    Statements Section    |
+---------------------------+
|     Exception Section     |
+---------------------------+
The Declaration section is the part defined between the PROCEDURE/FUNCTION header or the DECLARE keyword (for anonymous blocks) and the BEGIN keyword. (Optional section)
The Statements section is where your code goes and lies between the BEGIN keyword and the EXCEPTION keyword (or END keyword if there is no EXCEPTION section). (Mandatory section)
The Exception section is where any exception handling goes and lies between the EXCEPTION keyword at the END keyword. (Optional section)
Example of an anonymous block...
DECLARE
.. declarative statements go here ..
BEGIN
.. code statements go here ..
EXCEPTION
.. exception handlers go here ..
END;
Example of a procedure/function block...
[CREATE OR REPLACE] (PROCEDURE|FUNCTION) <proc or fn name> [(<parameters>)] [RETURN <datatype>] (IS|AS)
.. declarative statements go here ..
BEGIN
.. code statements go here ..
EXCEPTION
.. exception handlers go here ..
END;
(Note: The same can also be done for packages, but let's keep it simple)
2. Execution of the Execution Block
This may seem a simple concept, but it's surprising how many people have issues showing they haven't grasped it. When an Execution block is entered, the declaration section is processed, creating a scope of variables, types , cursors, etc. to be visible to the execution block and then execution enters into the Statements section. Each statment in the statements section is executed in turn and when the execution completes the last statment the execution block is exited back to whatever called it.
3. Exceptions
Exceptions generally happen during the execution of statements in the Statements section. When an exception happens the execution of statements jumps immediately into the exception section. In this section we can specify what exceptions we wish to 'capture' or 'trap' and do one of the two following things...
(Note: The exception section still has access to all the declared items in the declaration section)
3.i) Handle the exception
We do this when we recognise what the exception is (most likely it's something we expect to happen) and we have a means of dealing with it so that our application can continue on.
Example...
(without the exception handler the exception is passed back to the calling code, in this case SQL*Plus)
SQL> ed
Wrote file afiedt.buf
1 declare
2    v_name VARCHAR2(20);
3 begin
4    select ename
5    into   v_name
6    from   emp
7    where empno = &empno;
8    dbms_output.put_line(v_name);
9* end;
SQL> /
Enter value for empno: 123
old   7:   where empno = &empno;
new   7:   where empno = 123;
declare
ERROR at line 1:
ORA-01403: no data found
ORA-06512: at line 4
(with an exception handler, we capture the exception, handle it how we want to, and the calling code is happy that there is no error for it to report)
SQL> ed
Wrote file afiedt.buf
1 declare
2    v_name VARCHAR2(20);
3 begin
4    select ename
5    into   v_name
6    from   emp
7    where empno = &empno;
8    dbms_output.put_line(v_name);
9 exception
10    when no_data_found then
11      dbms_output.put_line('There is no employee with this employee number.');
12* end;
SQL> /
Enter value for empno: 123
old   7:   where empno = &empno;
new   7:   where empno = 123;
There is no employee with this employee number.
PL/SQL procedure successfully completed.
3.ii) Raise the exception
We do this when:-
a) we recognise the exception, handle it but still want to let the calling code know that it happened
b) we recognise the exception, wish to log it happened and then let the calling code deal with it
c) we don't recognise the exception and we want the calling code to deal with it
Example of b)
SQL> ed
Wrote file afiedt.buf
1 declare
2    v_name VARCHAR2(20);
3    v_empno NUMBER := &empno;
4 begin
5    select ename
6    into   v_name
7    from   emp
8    where empno = v_empno;
9    dbms_output.put_line(v_name);
10 EXCEPTION
11    WHEN no_data_found THEN
12      INSERT INTO sql_errors (txt)
13      VALUES ('Search for '||v_empno||' failed.');
14      COMMIT;
15      RAISE;
16* end;
SQL> /
Enter value for empno: 123
old   3:   v_empno NUMBER := &empno;
new   3:   v_empno NUMBER := 123;
declare
ERROR at line 1:
ORA-01403: no data found
ORA-06512: at line 15
SQL> select * from sql_errors;
TXT
Search for 123 failed.
SQL>
Example of c)
SQL> ed
Wrote file afiedt.buf
1 declare
2    v_name VARCHAR2(20);
3    v_empno NUMBER := &empno;
4 begin
5    select ename
6    into   v_name
7    from   emp
8    where empno = v_empno;
9    dbms_output.put_line(v_name);
10 EXCEPTION
11    WHEN no_data_found THEN
12      INSERT INTO sql_errors (txt)
13      VALUES ('Search for '||v_empno||' failed.');
14      COMMIT;
15      RAISE;
16    WHEN others THEN
17      RAISE;
18* end;
SQL> /
Enter value for empno: 'ABC'
old   3:   v_empno NUMBER := &empno;
new   3:   v_empno NUMBER := 'ABC';
declare
ERROR at line 1:
ORA-06502: PL/SQL: numeric or value error: character to number conversion error
ORA-06512: at line 3
SQL> select * from sql_errors;
TXT
Search for 123 failed.
SQL>
As you can see from the sql_errors log table, no log was written so the WHEN others exception was the exception that raised the error to the calling code (SQL*Plus)
4. Understanding Execution Blocks (part 2)
Ok, so now we understand the very basics of an execution block and what happens when an exception happens. Let's take it a step further...
Execution blocks are not just a single simple block in most cases. Often, during our statements section we have a need to call some reusable code and we do that by calling a procedure or function. Effectively this nests the procedure or function's code as another execution block within the current statement section so, in terms of execution, we end up with something like...
+---------------------------------+
|    Declaration Section          |
+---------------------------------+
|    Statements Section          |
|            .                    |
| +---------------------------+ |
| |    Declaration Section    | |
| +---------------------------+ |
| |    Statements Section    | |
| +---------------------------+ |
| |     Exception Section     | |
| +---------------------------+ |
|            .                    |
+---------------------------------+
|     Exception Section           |
+---------------------------------+
Example... (Note: log_trace just writes some text to a table for tracing)
SQL> create or replace procedure a as
2    v_dummy NUMBER := log_trace('Procedure A''s Declaration Section');
3 begin
4    v_dummy := log_trace('Procedure A''s Statement Section');
5    v_dummy := 1/0; -- cause an exception
6 exception
7    when others then
8      v_dummy := log_trace('Procedure A''s Exception Section');
9      raise;
10 end;
11 /
Procedure created.
SQL> create or replace procedure b as
2    v_dummy NUMBER := log_trace('Procedure B''s Declaration Section');
3 begin
4    v_dummy := log_trace('Procedure B''s Statement Section');
5    a; -- HERE the execution passes to the declare/statement/exception sections of A
6 exception
7    when others then
8      v_dummy := log_trace('Procedure B''s Exception Section');
9      raise;
10 end;
11 /
Procedure created.
SQL> exec b;
BEGIN b; END;
ERROR at line 1:
ORA-01476: divisor is equal to zero
ORA-06512: at "SCOTT.B", line 9
ORA-06512: at line 1
SQL> select * from code_trace;
TXT
Procedure B's Declaration Section
Procedure B's Statement Section
Procedure A's Declaration Section
Procedure A's Statement Section
Procedure A's Exception Section
Procedure B's Exception Section
6 rows selected.
SQL>
Likewise, execution blocks can be nested deeper and deeper.
5. How to continue exection of statements after an exception
One of the common questions asked is how to return execution to the statement after the one that created the exception and continue on.
Well, firstly, you can only do this for statements you expect to raise an exception, such as when you want to check if there is no data found in a query.
If you consider what's been shown above you could put any statement you expect to cause an exception inside it's own procedure or function with it's own exception section to handle the exception without raising it back to the calling code. However, the nature of procedures and functions is really to provide a means of re-using code, so if it's a statement you only use once it seems a little silly to go creating individual procedures for these.
Instead, you nest execution blocks directly, to give the same result as shown in the diagram at the start of part 4 of this article.
SQL> ed
Wrote file afiedt.buf
1 create or replace procedure b (p_empno IN VARCHAR2) as
2    v_dummy NUMBER := log_trace('Procedure B''s Declaration Section');
3 begin
4    v_dummy := log_trace('Procedure B''s Statement Section');
5    -- Here we start another execution block nested in the first one...
6    declare
7      v_dummy NUMBER := log_trace('Nested Block Declaration Section');
8    begin
9      v_dummy := log_trace('Nested Block Statement Section');
10      select empno
11        into   v_dummy
12        from   emp
13       where empno = p_empno; -- Note: the parameters and variables from
                                     parent execution block are available to use!
14    exception
15      when no_data_found then
16        -- This is an exception we can handle so we don't raise it
17        v_dummy := log_trace('No employee was found');
18        v_dummy := log_trace('Nested Block Exception Section - Exception Handled');
19      when others then
20        -- Other exceptions we can't handle so we raise them
21        v_dummy := log_trace('Nested Block Exception Section - Exception Raised');
22        raise;
23    end;
24    -- ...Here endeth the nested execution block
25    -- As the nested block handled it's exception we come back to here...
26    v_dummy := log_trace('Procedure B''s Statement Section Continued');
27 exception
28    when others then
29      -- We'll only get to here if an unhandled exception was raised
30      -- either in the nested block or in procedure b's statement section
31      v_dummy := log_trace('Procedure B''s Exception Section');
32      raise;
33* end;
SQL> /
Procedure created.
SQL> exec b(123);
PL/SQL procedure successfully completed.
SQL> select * from code_trace;
TXT
Procedure B's Declaration Section
Procedure B's Statement Section
Nested Block Declaration Section
Nested Block Statement Section
No employee was found
Nested Block Exception Section - Exception Handled
Procedure B's Statement Section Continued
7 rows selected.
SQL> truncate table code_trace;
Table truncated.
SQL> exec b('ABC');
BEGIN b('ABC'); END;
ERROR at line 1:
ORA-01722: invalid number
ORA-06512: at "SCOTT.B", line 32
ORA-06512: at line 1
SQL> select * from code_trace;
TXT
Procedure B's Declaration Section
Procedure B's Statement Section
Nested Block Declaration Section
Nested Block Statement Section
Nested Block Exception Section - Exception Raised
Procedure B's Exception Section
6 rows selected.
SQL>
You can see from this that, very simply, the code that we expected may have an exception was able to either handle the exception and return to the outer execution block to continue execution, or if an unexpected exception occurred then it was able to be raised up to the outer exception section.
6. User defined exceptions
There are three sorts of 'User Defined' exceptions. There are logical situations (e.g. business logic) where, for example, certain criteria are not met to complete a task, and there are existing Oracle errors that you wish to give a name to in order to capture them in the exception section. The third is raising your own exception messages with our own exception numbers. Let's look at the first one...
Let's say I have tables which detail stock availablility and reorder levels...
SQL> select * from reorder_level;
   ITEM_ID STOCK_LEVEL
         1          20
         2          20
         3          10
         4           2
         5           2
SQL> select * from stock;
   ITEM_ID ITEM_DESC STOCK_LEVEL
         1 Pencils             10
         2 Pens                 2
         3 Notepads            25
         4 Stapler              5
         5 Hole Punch           3
SQL>
Now, our Business has told the administrative clerk to check stock levels and re-order anything that is below the re-order level, but not to hold stock of more than 4 times the re-order level for any particular item. As an IT department we've been asked to put together an application that will automatically produce the re-order documents upon the clerks request and, because our company is so tight-ar*ed about money, they don't want to waste any paper with incorrect printouts so we have to ensure the clerk can't order things they shouldn't.
SQL> ed
Wrote file afiedt.buf
1 create or replace procedure re_order(p_item_id NUMBER, p_quantity NUMBER) is
2    cursor cur_stock_reorder is
3      select s.stock_level
4            ,r.stock_level as reorder_level
5            ,(r.stock_level*4) as reorder_limit
6      from stock s join reorder_level r on (s.item_id = r.item_id)
7      where s.item_id = p_item_id;
8    --
9    v_stock cur_stock_reorder%ROWTYPE;
10 begin
11    OPEN cur_stock_reorder;
12    FETCH cur_stock_reorder INTO v_stock;
13    IF cur_stock_reorder%NOTFOUND THEN
14      RAISE no_data_found;
15    END IF;
16    CLOSE cur_stock_reorder;
17    --
18    IF v_stock.stock_level >= v_stock.reorder_level THEN
19      -- Stock is not low enough to warrant an order
20      DBMS_OUTPUT.PUT_LINE('Stock has not reached re-order level yet!');
21    ELSE
22      IF v_stock.stock_level + p_quantity > v_stock.reorder_limit THEN
23        -- Required amount is over-ordering
24        DBMS_OUTPUT.PUT_LINE('Quantity specified is too much. Max for this item: '
                                 ||to_char(v_stock.reorder_limit-v_stock.stock_level));
25      ELSE
26        DBMS_OUTPUT.PUT_LINE('Order OK. Printing Order...');
27        -- Here goes our code to print the order
28      END IF;
29    END IF;
30    --
31 exception
32    WHEN no_data_found THEN
33      CLOSE cur_stock_reorder;
34      DBMS_OUTPUT.PUT_LINE('Invalid Item ID.');
35* end;
SQL> /
Procedure created.
SQL> exec re_order(10,100);
Invalid Item ID.
PL/SQL procedure successfully completed.
SQL> exec re_order(3,40);
Stock has not reached re-order level yet!
PL/SQL procedure successfully completed.
SQL> exec re_order(1,100);
Quantity specified is too much. Max for this item: 70
PL/SQL procedure successfully completed.
SQL> exec re_order(2,50);
Order OK. Printing Order...
PL/SQL procedure successfully completed.
SQL>
Ok, so that code works, but it's a bit messy with all those nested IF statements. Is there a cleaner way perhaps? Wouldn't it be nice if we could set up our own exceptions...
SQL> ed
Wrote file afiedt.buf
1 create or replace procedure re_order(p_item_id NUMBER, p_quantity NUMBER) is
2    cursor cur_stock_reorder is
3      select s.stock_level
4            ,r.stock_level as reorder_level
5            ,(r.stock_level*4) as reorder_limit
6      from stock s join reorder_level r on (s.item_id = r.item_id)
7      where s.item_id = p_item_id;
8    --
9    v_stock cur_stock_reorder%ROWTYPE;
10    --
11    -- Let's declare our own exceptions for business logic...
12    exc_not_warranted EXCEPTION;
13    exc_too_much      EXCEPTION;
14 begin
15    OPEN cur_stock_reorder;
16    FETCH cur_stock_reorder INTO v_stock;
17    IF cur_stock_reorder%NOTFOUND THEN
18      RAISE no_data_found;
19    END IF;
20    CLOSE cur_stock_reorder;
21    --
22    IF v_stock.stock_level >= v_stock.reorder_level THEN
23      -- Stock is not low enough to warrant an order
24      RAISE exc_not_warranted;
25    END IF;
26    --
27    IF v_stock.stock_level + p_quantity > v_stock.reorder_limit THEN
28      -- Required amount is over-ordering
29      RAISE exc_too_much;
30    END IF;
31    --
32    DBMS_OUTPUT.PUT_LINE('Order OK. Printing Order...');
33    -- Here goes our code to print the order
34    --
35 exception
36    WHEN no_data_found THEN
37      CLOSE cur_stock_reorder;
38      DBMS_OUTPUT.PUT_LINE('Invalid Item ID.');
39    WHEN exc_not_warranted THEN
40      DBMS_OUTPUT.PUT_LINE('Stock has not reached re-order level yet!');
41    WHEN exc_too_much THEN
42      DBMS_OUTPUT.PUT_LINE('Quantity specified is too much. Max for this item: '
                              ||to_char(v_stock.reorder_limit-v_stock.stock_level));
43* end;
SQL> /
Procedure created.
SQL> exec re_order(10,100);
Invalid Item ID.
PL/SQL procedure successfully completed.
SQL> exec re_order(3,40);
Stock has not reached re-order level yet!
PL/SQL procedure successfully completed.
SQL> exec re_order(1,100);
Quantity specified is too much. Max for this item: 70
PL/SQL procedure successfully completed.
SQL> exec re_order(2,50);
Order OK. Printing Order...
PL/SQL procedure successfully completed.
SQL>
That's better. And now we don't have to use all those nested IF statements and worry about it accidently getting to code that will print the order out as, once one of our user defined exceptions is raised, execution goes from the Statements section into the Exception section and all handling of errors is done in one place.
Now for the second sort of user defined exception...
A new requirement has come in from the Finance department who want to have details shown on the order that show a re-order 'indicator' based on the formula ((maximum allowed stock - current stock)/re-order quantity), so this needs calculating and passing to the report...
SQL> ed
Wrote file afiedt.buf
1 create or replace procedure re_order(p_item_id NUMBER, p_quantity NUMBER) is
2    cursor cur_stock_reorder is
3      select s.stock_level
4            ,r.stock_level as reorder_level
5            ,(r.stock_level*4) as reorder_limit
6            ,(((r.stock_level*4)-s.stock_level)/p_quantity) as finance_factor
7      from stock s join reorder_level r on (s.item_id = r.item_id)
8      where s.item_id = p_item_id;
9    --
10    v_stock cur_stock_reorder%ROWTYPE;
11    --
12    -- Let's declare our own exceptions for business logic...
13    exc_not_warranted EXCEPTION;
14    exc_too_much      EXCEPTION;
15 begin
16    OPEN cur_stock_reorder;
17    FETCH cur_stock_reorder INTO v_stock;
18    IF cur_stock_reorder%NOTFOUND THEN
19      RAISE no_data_found;
20    END IF;
21    CLOSE cur_stock_reorder;
22    --
23    IF v_stock.stock_level >= v_stock.reorder_level THEN
24      -- Stock is not low enough to warrant an order
25      RAISE exc_not_warranted;
26    END IF;
27    --
28    IF v_stock.stock_level + p_quantity > v_stock.reorder_limit THEN
29      -- Required amount is over-ordering
30      RAISE exc_too_much;
31    END IF;
32    --
33    DBMS_OUTPUT.PUT_LINE('Order OK. Printing Order...');
34    -- Here goes our code to print the order, passing the finance_factor
35    --
36 exception
37    WHEN no_data_found THEN
38      CLOSE cur_stock_reorder;
39      DBMS_OUTPUT.PUT_LINE('Invalid Item ID.');
40    WHEN exc_not_warranted THEN
41      DBMS_OUTPUT.PUT_LINE('Stock has not reached re-order level yet!');
42    WHEN exc_too_much THEN
43      DBMS_OUTPUT.PUT_LINE('Quantity specified is too much. Max for this item: '
                              ||to_char(v_stock.reorder_limit-v_stock.stock_level));
44* end;
SQL> /
Procedure created.
SQL> exec re_order(2,40);
Order OK. Printing Order...
PL/SQL procedure successfully completed.
SQL> exec re_order(2,0);
BEGIN re_order(2,0); END;
ERROR at line 1:
ORA-01476: divisor is equal to zero
ORA-06512: at "SCOTT.RE_ORDER", line 17
ORA-06512: at line 1
SQL>
Hmm, there's a problem if the person specifies a re-order quantity of zero. It raises an unhandled exception.
Well, we could put a condition/check into our code to make sure the parameter is not zero, but again we would be wrapping our code in an IF statement and not dealing with the exception in the exception handler.
We could do as we did before and just include a simple IF statement to check the value and raise our own user defined exception but, in this instance the error is standard Oracle error (ORA-01476) so we should be able to capture it inside the exception handler anyway... however...
EXCEPTION
WHEN ORA-01476 THEN
... is not valid. What we need is to give this Oracle error a name.
This is done by declaring a user defined exception as we did before and then associating that name with the error number using the PRAGMA EXCEPTION_INIT statement in the declaration section.
SQL> ed
Wrote file afiedt.buf
1 create or replace procedure re_order(p_item_id NUMBER, p_quantity NUMBER) is
2    cursor cur_stock_reorder is
3      select s.stock_level
4            ,r.stock_level as reorder_level
5            ,(r.stock_level*4) as reorder_limit
6            ,(((r.stock_level*4)-s.stock_level)/p_quantity) as finance_factor
7      from stock s join reorder_level r on (s.item_id = r.item_id)
8      where s.item_id = p_item_id;
9    --
10    v_stock cur_stock_reorder%ROWTYPE;
11    --
12    -- Let's declare our own exceptions for business logic...
13    exc_not_warranted EXCEPTION;
14    exc_too_much      EXCEPTION;
15    --
16    exc_zero_quantity EXCEPTION;
17    PRAGMA EXCEPTION_INIT(exc_zero_quantity, -1476);
18 begin
19    OPEN cur_stock_reorder;
20    FETCH cur_stock_reorder INTO v_stock;
21    IF cur_stock_reorder%NOTFOUND THEN
22      RAISE no_data_found;
23    END IF;
24    CLOSE cur_stock_reorder;
25    --
26    IF v_stock.stock_level >= v_stock.reorder_level THEN
27      -- Stock is not low enough to warrant an order
28      RAISE exc_not_warranted;
29    END IF;
30    --
31    IF v_stock.stock_level + p_quantity > v_stock.reorder_limit THEN
32      -- Required amount is over-ordering
33      RAISE exc_too_much;
34    END IF;
35    --
36    DBMS_OUTPUT.PUT_LINE('Order OK. Printing Order...');
37    -- Here goes our code to print the order, passing the finance_factor
38    --
39 exception
40    WHEN exc_zero_quantity THEN
41      DBMS_OUTPUT.PUT_LINE('Quantity of 0 (zero) is invalid.');
42    WHEN no_data_found THEN
43      CLOSE cur_stock_reorder;
44      DBMS_OUTPUT.PUT_LINE('Invalid Item ID.');
45    WHEN exc_not_warranted THEN
46      DBMS_OUTPUT.PUT_LINE('Stock has not reached re-order level yet!');
47    WHEN exc_too_much THEN
48      DBMS_OUTPUT.PUT_LINE('Quantity specified is too much. Max for this item: '
                              ||to_char(v_stock.reorder_limit-v_stock.stock_level));
49* end;
SQL> /
Procedure created.
SQL> exec re_order(2,0);
Quantity of 0 (zero) is invalid.
PL/SQL procedure successfully completed.
SQL>
Lastly, let's look at raising our own exceptions with our own exception numbers...
SQL> ed
Wrote file afiedt.buf
1 create or replace procedure re_order(p_item_id NUMBER, p_quantity NUMBER) is
2    cursor cur_stock_reorder is
3      select s.stock_level
4            ,r.stock_level as reorder_level
5            ,(r.stock_level*4) as reorder_limit
6            ,(((r.stock_level*4)-s.stock_level)/p_quantity) as finance_factor
7      from stock s join reorder_level r on (s.item_id = r.item_id)
8      where s.item_id = p_item_id;
9    --
10    v_stock cur_stock_reorder%ROWTYPE;
11    --
12    exc_zero_quantity EXCEPTION;
13    PRAGMA EXCEPTION_INIT(exc_zero_quantity, -1476);
14 begin
15    OPEN cur_stock_reorder;
16    FETCH cur_stock_reorder INTO v_stock;
17    IF cur_stock_reorder%NOTFOUND THEN
18      RAISE no_data_found;
19    END IF;
20    CLOSE cur_stock_reorder;
21    --
22    IF v_stock.stock_level >= v_stock.reorder_level THEN
23      -- Stock is not low enough to warrant an order
24      [b]RAISE_APPLICATION_ERROR(-20000, 'Stock has not reached re-order level yet!');[/b]
25    END IF;
26    --
27    IF v_stock.stock_level + p_quantity > v_stock.reorder_limit THEN
28      -- Required amount is over-ordering
29

its nice article, have put up this one the blog
site,Nah, I don't have time to blog, but if one of the other Ace's/Experts wants to copy it to a blog with reference back to here (and all due credit given ;)) then that's fine by me.
I'd go for a book like "Selected articles by OTN members" or something. Does anybody have a list of links of all those mentioned articles?Just these ones I've bookmarked...
Introduction to regular expressions ... by CD
When your query takes too long ... by Rob van Wijk
How to pipeline a function with a dynamic number of columns? by ascheffer
PL/SQL 101 : Exception Handling by BluShadow

Exception Handling-rite way??

Hi Friends,
This Exception handling is really causing some problems for me.I run a query,return the resultset,cook the data from my other java class and display it thru my jsp and the last statement from my jsp is to call the close method(commented out).The problem is if some unknown Exception arises the close() method is not being called,causing open connections which later on are
creating some disasters.I tried to implement it now using the finally method,so that it always gets closed,but hte problem is when i call the ReturnResultSet() method and try to cook the data,it says "ResultSet Closed".Please tell me which is the right way to implement this:
public ResultSet ReturnResultSet(String Query) throws Exception
     try{
       if (datasource != null) {
         connection = datasource.getConnection();
         if (connection != null) {
           statement = connection.createStatement( );
           resultset = statement.executeQuery(Query);
       return resultset;
     } catch (SQLException sqle)
       sqle.printStackTrace();
       return null;
     finally {
     try {
       if (resultset != null) resultset.close();
       if (statement != null) statement.close();
       if (connection != null) connection.close();
     catch (SQLException sqle) {
       sqle.printStackTrace();
public void close()
   try { resultset.close(); } catch (Exception ex) {}
   try { statement.close(); } catch (Exception ex) {}
   try { connection.close(); } catch (Exception ex) {}
*/Any help would be appreciated and some duke dollars would be awarded too.Thanks

Ok I think i got your point and i should award you
the duke dollars too,but one last thing to ask.I call
the close() method after all my processing is over,I
just
wanna know should I have the connection.close() thing
inside it,becuase dont that contradicts the whole
connection pool thing,as i am closing a connection
and it has to open a new connection for every
request.Or should i just have resultset.close() and
statement.close() in it.
Thanks for all your helpAre you talking about a standard J2EE container-provided connection pool? If so, then yes, you still need to 'close' the connection. That doesn't actually close it, it just tells the pool it is available to be used again the next time someone asks it for a connection. Hopefully you're not writing your own home-grown "connection pool".

Multiple Idocs to XI with exception handling.

Hi,
I m doing a scenario,sending multiple idocs to file via XI.while doing dis,i m using BPM.For the BPM part i m following the blog https://www.sdn.sap.com/irj/sdn/weblogs?blog=/pub/wlg/2034..but [original link is broken] [original link is broken] [original link is broken] [original link is broken] wen i check in SXMB_MONI,my message isn't successfully processed out f BPM.In this,exception is triggered at d beginning.and it always throw exception and message not successfully processed out f BPM.In the workflow it is showing error in the first step and remaining steps are compIeted.Here, i m throwing exception by the use f deadline branch.Can,anybody tell me how to handle dis exception/error?

Hi Rajeev,
I am using a loop with in a block inside the BPM inside which receive step and container element is used.In parallel to loop branch,i m using deadline and exception branch same as in blog 2034.i m using infinite loop wid time limit mentioned in the deadline branch.After coming out f dis block,i m using transformation step and den send step is used.When i check in graphical workflow,i m getting error in send step stating that an empty container element is used while sending.But,wen i m checking technical workflow,i m getting error in first step after the creation of workflow and workflow started messages in the log details.after which the exception occured is shown in the workflow system..
Thanks,
Anoop Garg

Flow Activity Exception Handling Problem

Hi,
I am using Flow Activity in my Bpel Process, i am getting some exception in one of flow, but other flows are working fine. Even i have included Catch block for other flow still i am getting fault response.
Please suggest me how to do exception handling in Flow Activity.
Thanks in advance.

Hi,
just restructure your BPEL process. The 'Flow' activity contains several 'Sequence' activities. Simply ... put a 'Scope' activity inside of each 'Sequence' activity. Then put another 'Sequence' activity inside of each 'Scope' activity ... and put there required logic/activities.
After that you can create fault handling on these 'Scope' activities. In this way you can handle faults in each flow-sequence.
Regards,
Martin

RFC: Exception handling

Hi there,
after playing a bit with external xml files provided by my
application i'm ready to go to the next topic on my ToDo-List.
How could one implement a good and stable technique for
exception handling? I'm talking about the interaction from spry to
our Web Framework which throws an error sometimes.
My current idea is that the server side framework could
return some kind of Error XML format which contains the exception
message and so on. This would require functions on client side to
precheck the server output.
Imho the best place for this is inside the XMLDataSet
constructor. When catching some bad XML file internal error
messages could be send using the already existing techniques OR by
providing an public function which then should be overriden by an
selfwritten one, e.g. to put it into my own `error box`.
Also, some kind of internal statemachine in XMLDataSet may
help a lot. That way each access to functions which aren't working
at the moment could be loggend and catched by an selfwritten
function.
I hope you get the idea,
what do you think?
Best regards,
Sebastian

Better error handling and error handling hooks is on the list
of things to do.
The way I see it, there are several types of errors that can
occur:
1. Server returns valid XML, but it's XML that describes an
error instead of the data requested.
- I believe this is what Sebastian was mentioning. I was
actually thinking of allowing a hook for developers to catch and
handle this case and perhaps leverage the states mechanism to let
them change the dynamic region markup used to display the error
since the data references in this error XML would be different.
2. The server returns an error. (Invalid URL or Server Error)
- This could be handled with states, but we need to expose
some data references, or set the data set to contain a known data
set schema that would allow the designer to show more info about
the error.
3. The server returns XML but uses a mime-type that is not
understood by Spy or the XML parsing code built-into the browser.
- I believe it was Doug [?] that had a patch that *always*
forced the data set to try and parse the XML string in the response
if the response didn't contain an XML DOM. My one paranoia about
that is that the server could actually be returning something that
is not XML, in which we would still fail and perhaps choke
somewhere else. I need to do some testing in that area.
I was thinking perhaps we should add something to the
XMLDataSet constructor that allowed a user to specify mime-types
for formats they knew were XML, but didn't use one of the standard
XML formats.
4. The browser chokes on "not-well-formed" XML.
- This is an interesting problem. IE silently fails when the
parser chokes, but Mozilla creates an XML DOM tree that reports the
error which does *not* match the XML string from the request
response. I had to add code to spry to detect when this happens.
5. An exception is thrown during Spry processing of XML data.
- This will require more programming on our part to handle
more cases.
--== Kin ==--

JSF Exception Handling Example

Here is an example of catching a backend exception and displaying it on the UI. It seems like most of the examples in books and such concern themselves with Validators and Converters. If you are trying to catch a backend exception, and you just want to display a UI message for it, see if this will work for you:
JSF tags:
<h:commandLink id="myActionLink" value="#{msgs.myLinkText}" binding="#{MyFormBean.uiComponentLink}" action="#{MyFormBean.linkProcessingMethod}"></h:commandLink>
<h:messages layout="table" errorClass="Errors" styleClass="AlignLeft"/>
...And now the exception handling and message creation inside of the managed bean:
private HtmlCommandLink uiComponentLink;
* Method that processes the link action
* @return String (action ID)
public String linkProcessingMethod()
     try
          // BEGIN: TEST ONLY
          if(1 == 1)     // Always true
               throw new Exception("TEST ERROR MESSAGE DISPLAY");
          // END: TEST ONLY
     return "dummyPageReference";     // No reference for this in faces-config.xml. Redisplays same page.
     catch (Exception e)
         addFacesMessageForUI("Exception detected. Message: " +e.getMessage());
         return "dummyPageReference";
* Add faces message for display on the UI
* @param String uiMessage
private void addFacesMessageForUI(String uiMessage)
     FacesMessage facesMessage = new FacesMessage(uiMessage);
     FacesContext facesContext = FacesContext.getCurrentInstance();
     // Passing null for the client ID argument to the FacesContext
     // addMessage() method specifies the message as not belonging
     // to any particular UI component. This will cause the message
     // to be displayed as a general message on the UI
     facesContext.addMessage(null,facesMessage);
}

chrisjohn wrote:
Following link would be helpful
[http://technologicalbrainstorm.wordpress.com/2009/09/19/exception-handling-in-jsf/]
Thanks!!Please, don't resurrect old threads, and especially not to post link-spam. Your account will get blocked if you continue with it. I'm locking this thread.
Kaj

Add three stored procs in one big proc and add exception handling

I have three proc's,
1. sp_staging
2. sp_upload
3. sp_process.
Now, I have to create a new proc as sp_daily which has the other three proc's in it in the same order.
if the first proc sp_staging fails, then the sp_daily proc should stop executing the other proc's which are in the order.
CREATE PROCEDURE sp_daily
AS
BEGIN TRY
EXEC sp_staging (already has EXCEPTION handling IN it)
GO
EXEC sp_upload (already has EXCEPTION handling IN it)
GO
EXEC sp_process (already has EXCEPTION handling IN it)
GO
END TRY
BEGIN CATCH
-- If any of the above proc's fail, the next proc's should not be executed
END CATCH
Something like this????

Hi naveej,
Whether the Stored Procedure sp_daily stops executing if the first sp_staging fails depends on how you code. To give a better demonstration, I will show the scenario that sp_upload runs with error. Please see the below code.
--create the scenario
USE master;
IF db_id('TestDB') IS NULL
CREATE DATABASE TestDB;
USE TestDB;
IF OBJECT_ID('T1') IS NOT NULL
DROP TABLE T1;
GO
CREATE TABLE T1
procName varchar(99)
IF OBJECT_ID('sp_staging') IS NOT NULL
DROP PROCEDURE "sp_staging";
GO
CREATE PROCEDURE "sp_staging" @denominator FLOAT
AS
BEGIN TRY
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator; -- when @denominator=0, error occurs
INSERT INTO T1 VALUES('sp_staging finished');
END TRY
BEGIN CATCH
END CATCH
GO
IF OBJECT_ID('sp_upload') IS NOT NULL
DROP PROCEDURE "sp_upload";
GO
CREATE PROCEDURE "sp_upload" @denominator FLOAT
AS
BEGIN TRY
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator;
INSERT INTO T1 VALUES('sp_upload finished');
END TRY
BEGIN CATCH
DECLARE @ERRMSG VARCHAR(99);
SELECT @ERRMSG=ERROR_MESSAGE()+'sp_upload failed, process continues';
INSERT INTO T1 VALUES(@ERRMSG);
END CATCH
GO
IF OBJECT_ID('sp_process') IS NOT NULL
DROP PROCEDURE "sp_process";
GO
CREATE PROCEDURE "sp_process" @denominator FLOAT
AS
BEGIN TRY
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator;
INSERT INTO T1 VALUES('sp_process finished');
END TRY
BEGIN CATCH
END CATCH
GO
IF OBJECT_ID('sp_daily') IS NOT NULL
DROP PROCEDURE "sp_daily";
GO
CREATE PROCEDURE "sp_daily" @p1Deno FLOAT,@p2Deno FLOAT,@p3Deno FLOAT
AS
BEGIN TRY
EXEC sp_staging @p1Deno;
EXEC sp_upload @p2Deno;
EXEC sp_process @p3Deno;
END TRY
BEGIN CATCH
INSERT INTO T1 SELECT ERROR_MESSAGE()+'error captured by sp_daily';
END CATCH
GO
--Test Example
TRUNCATE TABLE T1;
EXEC sp_daily 1,0,1;
SELECT * FROM T1;
We can judge from the rows in T1, sp_daily continued even when there came a error in sp_upload. The reason for the continuity is that the error which occurred in sp_upload was caught and handled by sp_upload itself. Sp_daily's CATCH block didn't capture
any error(we didnt see any row like "error captured by sp_daily") so sp_dailys continued running with no interruption.
Let’s make a little modification on the sp_upload. see the below code.
ALTER PROCEDURE "sp_upload" @denominator FLOAT
AS
BEGIN TRY
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator;
INSERT INTO T1 VALUES('sp_upload');
END TRY
BEGIN CATCH
DECLARE @ERRMSG VARCHAR(99),@ErrSeverity INT;
SELECT @ERRMSG=ERROR_MESSAGE()+'sp_upload failed, process got terminated',@ErrSeverity=ERROR_SEVERITY();
RAISERROR(@ERRMSG ,@ErrSeverity,1);
END CATCH
GO
--Test Example
TRUNCATE TABLE T1;
EXEC sp_daily 1,0,1;
SELECT * FROM T1;
At this time, sp_daily got terminated when sp_upload encountered an error. Sp_upload didn’t handle the error, it threw the error instead. Then the CATCH block in sp_daily caught the error and interrupted sp_daily itself. If you don’t
need to handle the error which may occur inside of the inner called procedures, the TRY and CATCH block in the procedure which calls them would be enough to catch and handle the errors anywhere it occurs, just remove the blocks from the called ones.
ALTER PROCEDURE "sp_staging" @denominator FLOAT
AS
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator;
INSERT INTO T1 VALUES('sp_staging finished');
GO
ALTER PROCEDURE "sp_upload" @denominator FLOAT
AS
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator;
INSERT INTO T1 VALUES('sp_upload finished');
GO
ALTER PROCEDURE "sp_process" @denominator FLOAT
AS
DECLARE @quotient FLOAT;
SET @denominator=1/@denominator;
INSERT INTO T1 VALUES('sp_process finished');
GO
ALTER PROCEDURE "sp_daily" @p1Deno FLOAT,@p2Deno FLOAT,@p3Deno FLOAT
AS
BEGIN TRY
EXEC sp_staging @p1Deno;
EXEC sp_upload @p2Deno;
EXEC sp_process @p3Deno;
END TRY
BEGIN CATCH
INSERT INTO T1 SELECT ERROR_MESSAGE();
END CATCH
GO
--Test Example
TRUNCATE TABLE T1;
EXEC sp_daily 1,0,1;
SELECT * FROM T1;
As we can see from the above modification, sp_daily calls sp_staging, sp_upload, sp_process in order. Sp_staging runs fine so it finishes. When it comes to sp_upload, error gets captured so the execution terminates. Anyway if you hope that the 3
called procedures work atomically(it is a pretty common business requirement), which means unless all of the three finish otherwise no one finishes, the TRAN block would help. Please see the below code.
ALTER PROCEDURE "sp_daily" @p1Deno FLOAT,@p2Deno FLOAT,@p3Deno FLOAT
AS
BEGIN TRY
BEGIN TRAN
EXEC sp_staging @p1Deno;
EXEC sp_upload @p2Deno;
EXEC sp_process @p3Deno;
COMMIT TRAN;
END TRY
BEGIN CATCH
ROLLBACK TRAN;
INSERT INTO T1 SELECT ERROR_MESSAGE();
END CATCH
GO
--Test Example
TRUNCATE TABLE T1;
EXEC sp_daily 1,0,1;
SELECT * FROM T1;
If you have question, feel free to let me know.
Best Regards,
Eric Zhang

Exception handling is not working in GCC compile shared object

Hello,
I am facing very strange issue on Solaris x86_64 platform with C++ code compiled usging gcc.3.4.3.
I have compiled shared object that load into web server process space while initialization. Whenever any exception generate in code base, it is not being caught by exception handler. Even though exception handlers are there. Same code is working fine since long time but on Solaris x86, Sparc arch, Linux platform
With Dbx, I am getting following stack trace.
Stack trace is
dbx: internal error: reference through NULL pointer at line 973 in file symbol.cc
[1] 0x11335(0x1, 0x1, 0x474e5543432b2b00, 0x59cb60, 0xfffffd7fffdff2b0, 0x11335), at 0x11335
---- hidden frames, use 'where -h' to see them all ----
=>[4] __cxa_throw(obj = (nil), tinfo = (nil), dest = (nil), , line 75 in "eh_throw.cc"
[5] OBWebGate_Authent(r = 0xfffffd7fff3fb300), line 86 in "apache.cpp"
[6] ap_run_post_config(0x0, 0x0, 0x0, 0x0, 0x0, 0x0), at 0x444624
[7] main(0x0, 0x0, 0x0, 0x0, 0x0, 0x0), at 0x42c39a
I am using following link options.
Compile option is
/usr/sfw/bin/g++ -c -I/scratch/ashishas/view_storage/build/coreid1014/palantir/apache22/solaris-x86_64/include -m64 -fPIC -D_REENTRANT -Wall -g -o apache.o apache.cpp
Link option is
/usr/sfw/bin/g++ -shared -m64 -o apache.so apache.o -lsocket -lnsl -ldl -lpthread -lthread
At line 86, we are just throwing simple exception which have catch handlers in place. Also we do have catch(...) handler as well.
Surpursing things are..same issue didn't observe if we make it as executable.
Issue only comes if this is shared object loaded on webserver. If this is plain shared object, opened by anyother exe, it works fine.
Can someone help me out. This is completly blocking issue for us. Using Solaris Sun Studio compiler is no option as of now.

shared object that load into web server process space
... same issue didn't observe if we make it as executable.When you "inject" your shared object into some other process a well-being of your exception handling depends on that other process.
Mechanics of x64 stack traversing (unwind) performed when you throw the exception is quite complicated,
particularly involving a "nearly-standartized" Unwind interface (say, Unwind_RaiseException).
When we are talking about g++ on Solaris there are two implementations of unwind interface, one in libc and one in libgcc_s.so.
When you g++-compile the executable you get it directly linked with libgcc_s.so and Unwind stuff resolves into libgccs.
When g++-compiled shared object is loaded into non-g++-compiled executable's process _Unwind calls are most likely already resolved into Solaris libc.
Thats why you might see the difference.
Now, what exactly causes this difference can vary, I can only speculate.
All that would not be a problem if _Unwind interface was completely standartized and properly implemented.
However there are two issues currently:
* gcc (libstdc++ in particular) happens to use additional non-standard _Unwind calls which are not present in Solaris libc
naturally, implementation details of Unwind implementation in libc differs to that of libgccs, so when all the standard _Unwind
routines are resolved into Solaris version and one non-standard _Unwind routine is resolved into gcc version you get a problem
(most likely that is what happens with you)
* libc Unwind sometimes is unable to decipher the code generated by gcc.
However that is likely to happen with modern gcc (say, 4.4+) and not that likely with 3.4.3
Btw, you can check your call frame to see where _Unwind calls come from:
where -h -lIf you indeed stomped on "mixed _Unwind" problem then the only chance for you is to play with linker
so it binds Unwind stuff from your library directly into libgccs.
Not tried it myself though.
regards,
__Fedor.

Exception Handling of Screenflows inside Processes

Similar Messages

Maybe you are looking for