Thanks all for your comments

as long as the point is clear - mixing TRY with @@error can cause subtle problems

and this point is not made in BOL, breaking changes, etc.

and...

Again, take a deep breath. You're not thinking. Everyone is telling you the same thing and you're not listening. If a higher severity error is unhanded in...

that is what I *am* doing! X and Y are...

The Dixie Flatline (7/7/2010)In your example, change the error severity to 9 instead of 16 and you get the consistent behavior you are after. The various behaviors...

Nils Gustav Stråbø (7/7/2010)

---

I guess [field] is a datetime column. Then you can rewrite your WHERE clause

WHERE [field]< {dateadd formula here}

Putting your SARGable columns in functions (like datediff, substring,...

I'm sorry to have to say it - but you still do not understand

OO

Grant Fritchey (7/7/2010)

---

You're dealing with a fundamental change in behavior. If you're raising errors inside the procedure, the errors are recognized by the TRY/CATCH construct, as designed. That's how it's...

Grant Fritchey (7/7/2010)

---

But if you have pre-existing procs with established and functioning logic, why are you changing them to use TRY/CATCH at all? I would only suggest rewriting them as...

Mr_Bacon (7/7/2010)Doobya, thanks for your feedback. Just for my info what do you mean when you say ‘you can't use datediff() like that - it will force a table scan’...

Matt Miller (#4) (7/6/2010)I do find it has some flaws, but still - it seems to be to be a huge step forward from continuously having to check @@ERROR after...

the problem is that it changes the behaviour of OTHER pre-existing stored procedures

that already have a defined and tested behaviour

it breaks them

a simple example:

if @@error <> 0 goto label_do_something_else

won't work...

if you are trying to achieve this:

Year, Course, Year, Course

08/09 Math09/10 English

08/09 Physics09/10 Math

08/09 Chemistry

I think you are trying to do a layout job in SQL ... that isn't right

return...

when you say "not updated" ... do you mean as in:

update table

set [field] = 'value'

where 1 = 0

type of not updated?

update stuff

if @@rowcount < 1 raiserror('oh no', 16, 1)

alter database set restricted_user

I always wished for:

alter database set single_restricted_user

for 5

5 * 4 * 3 * 2 * 1 = 120

for 4

5 * 4 * 3 * 2 = 120

for 3

5 * 4 * 3 = 60

for 2

5 *...