why do people always want to use cursors?

Embrace relational theory:

UPDATE f

   SET close_date = w.maxdate

  FROM feeback f

       JOIN (

              select feedbackid, MAX( response_date ) AS maxdate

                from responses GROUP...

default constraints are in sysobjects type = 'D'

values for these are in syscomments where id = sysobjects id

table object id (sysobjects) is in the parent_obj column of the constraint row...

you can use various means using

CASE ...look it up in online help

RIGHT( '000' + RTRIM( numbercolumn ) , 3 )

and/or

PATINDEX( '%[A-Z]%, numbercolumn)

(THAT IS, I don't know what possibilities...

um, no, and I never beat my wife either.

there's nothing special about the dash (or any other character) unless it is stripped by your client process on loading.  So take...

from what you wrote, it looks like you need varchar(20) and not (10)

it truncates at the tenth char and these two would then be identical

If you really have no options for space, and don't want to BCP the [desired] data out, truncate, and BCP back in (you know BCP supports select statements now)...

I had...

And to do that...

you can right a proc that you run once a minute (or whatever) to look at the sysprocesses table for SPID, login_date, sid, and what ever else...

I'm not having trouble with the norecovery option on sql2k. The upgrade process only happens when the 'with recovery' option is selected. What's the syntax you're using?

See the sql help on OBJECTPROPERTY function.

The first question has got to be, why do you need a cursor?

but apart from that...

why do you not just load the data from the spreadsheet into a temp or...

If this job runs every day and needs the extra space then why do you want to shrink the DB (and force it to have to reallocate the extra space...

The last_batch column is the date/time of the prior statement.

Probably this in conjunction with cmd = "AWAITING COMMAND" is sufficient (you don't want to kill something currently involved in a...

it sounds like you mean that it is a two part primary key dp + num.

so then you could do exactly the same thing except with a self join

select...

FROM legal...

yes but the temp table layout must match what the proc is going to spit out:

create proc itest

as select id, name from sysobjects

GO

--1 create temp table

select id, name into #t...

it sounds like it is behaving as if the "NORECOVERY" option is set (meaning waiting for further tansaction logs to be applied.)

try it the old fashioned way by TSQL:

RESTORE DATABASE...