yep the issue is what i...

if this is updating a table in a different database, it might be OK...

i see it is updating renovest.dbo.[Testtable$Afsl_ Vejehoved] ...what database is the table with the trigger in? drew's...

awesome!

like i said, it worked perfect for me with tables, but most of my views and procs failed; they all pointed to dbo.Tables....

i cleaned them all up for a single...

Howard just finding rows with lots of underscores, or are the underscores a placeholder for variable length data?

this finds them easily, for example:

With mySampleData(TheField)

AS

(

SELECT '20556116 29 _________________________________________________ 49' UNION ALL

SELECT...

I'd always tackled this the other way...backup and restore to the dev database, then DELETE any data that doesn't match the saving criteria...works great if your FK's have the DELETE...

is the slow version running a different SQL version?

it's very common for a database restored on a higher version of SQL to require updating statistics.

update the statistics on the table...

opc.three i absolutely agree; i just happen to have a saved snippet that did what the OP was asking for...i was thinking of re-doing it, since it doesn't cover shemas;...

i'd used this in the past...i originally made it in SQL 2000, so it pretty much ignores schemas and assumes dbo.

see how it;s using a cursor to loop through what...

i think you answered this in your first post...but just in case...

can you go to this spot and put sql files in there? does that make any difference?

i know...

show us your trigger, and we can help you tweak it to handle multiple rows as a set based solution..

as far as whether it's insert/update/delete, if there is any data...

there is a view you can use, that is mostly correct:

SELECT OBJECT_NAME(REFERENCING_ID),* FROM sys.sql_expression_dependencies

WHERE OBJECT_NAME(REFERENCING_ID = 'YourProcName'

issues can arise sometimes, becasue you can create a view that references tables...

definitely the conversion in there: here's a stark example...you'll see none of the results are the same,all due to the conversion issues:

/*

(No column name)(No column name)(No column name)(No column name)(No...

ok, i can duplicate some of what you are seeing...

try this, which is barely modified form what you posted....

i get chinese characters from my loginnmae due to the conversion:

what i...

as far as I know, the data type and sizes should always be varvbinary 128 and nvarchar(64), right?

wouldn't an implcit comversion form nvarchar to varchar mess up the data?

or is...

you can do what you want with the dateadd/datediff functions

In this case, one easy way to do it is to find the beginning of the month for the date in...