Something like this may work

log insert - beginning sp

begin tran

--table1 insert

if errored

rollback

goto err1

--table2 insert

if errored

rollback

goto err2

--table3 insert

if errored

rollback

goto...

One workaround would be to change the column to an int or smallint.

Is it a clustered index? Is there a clustered index on the table? Are there any other...

Does that option even have an effect in SQL2k? I thought it was already in ram in versions > 6.5?

quote:

---

(Shame all the formatting / tabbing will be lost when posted)

---

I had the same issue...

tsql help -> save transaction -> transaction savepoints

would be one way of keeping your logging in the same transaction yet allow you to roll back what didn't work. I...

I'm unclear on what results from including more or less parms. I am assuming there is more than one reason for having them (and hence your question regarding nulls),...

Ok.. so im slow hehe

I usually specifically keep any logging actions I am doing outside of transactions - if the transaction is rolled back you lose your log of what was going on.

Aye noel's will work better for you i'm sure.

I am in the habit of using 'in-line updates' a lot because it seems to suit my needs well (dramtically different from...

quote:

---

I'm using it to get a trace functionality similar to that of ASP.NET, but it uses OutputDebugString via an extended stored procedure...

 

drop table #SourceTable

go

-- building a table to take the place of your source table

create table #SourceTable (

 idsItem int not null

,idsParent int not null

,intIndex int...

not sure if it will speed it up or not but

(clm.settle_date < @todaysDate OR clm.settle_date is null)

could be

isnull(clm.settle_date,@TomorrowsDate) < @todaysDate)

Unless you have a ton of SQL servers and don't know which one(s) it's using, SQL profiler may be helpful.

It has the capacity to record/show all SQL statements issued. ...

Would using char/varchar but having a check on the field to ensure numeric(3,2) validity help?

Are you running any linked server queries to a non-server datasource, like foxpro or something?

If that 128mb(?) of memory it allows gets used up by multi-user or ineffecient table scans...