Reputation: 127
Is there a way to handle errors in SYBASE
, such as the TRY-CATCH
block you can use in MS SQL Server
, Oracle
, etc?
I've searched the web and the only option I found was the global variable @@error
, but it didn' work as I expected, for example, the following code:
begin tran
update table1
set name = 'new name'
where name = 'old name'
update table2
set id = 1
where id = 30
-- suppose id has a unique constraint and there's already a row with id = 1
IF @@error = 0
begin
print 'commited'
commit
end
else
begin
print 'rolled back'
rollback
end
The will indeed rollback somehow, because the name I've changed on table1 keeps the old value as I've tested here, but it doesn't print the messages, or execute any instructions I put after the instructions that causes the error
Can anyone help me in this? Do you know how does Sybase error handling actually works?
Upvotes: 4
Views: 24096
Reputation: 25753
1st solution.
You can't catch an exception this way on Sybase. Before you update you have to check data:
if not exists
(
select 1 from table2
where id = 1
)
begin
update table2
set id = 1
where id = 30
end
else
begin
print 'rolled back'
rollback
end
2nd solution.
You can also put an update command to procedure, then you can catch an exception. Create procedure:
create procedure myproc
as
begin
update table2
set id = 1
where id = 30
end
and run it as below:
begin tran
update table1
set name = 'new name'
where name = 'old name'
exec myproc
IF @@error = 0
begin
print 'commited'
commit
end
else
begin
print 'rolled back'
rollback
end
Upvotes: 3