I tried running your earlier code and from how it behaves I think you are not really interested in random numbers. I think what you want is sequential numbers for...

Oops! Isn't this a triangular join? Jeff will pounce on you for this.

Michael Meierruth (8/20/2008)

I have forgiven myself for not knowing the meaning of that abbreviation. But the same...

I have forgiven myself for not knowing the meaning of that abbreviation. But the same has not...

George,

Glad to hear this good news.

So what was your recovery mode setting?

Surely selecting the good records (2908 out of 6M) into a table, truncating and going back would have been...

If your transaction log is 'going to blow' it will do so with or without the delete crawler. Nevertheless, check your db recovery mode setting.

First of all, the delete statement I posted has a slight problem. It works correctly for the duplicates and would delete the duplicates and leave only one row with the...

Have a go with this.

delete from AssociationLinks

where RecordId not in

(

select min(RecordId)

from AssociationLinks

group by WorkspaceEntityId,SourceEntityId,TargetEntityId,Name,RecordStatus

having count(*)>1

)

I tried this with a...

Why not just keep looping through the tables. Eventually the constraints will go away.

Your might want to use truncate instead of delete.

When I run this, it returns 2008-05-04 which isn't even a Sunday and which with respect to today 's date(2008-05-16) should return 2008-05-11.

If you don't want to install the Oracle client on the SQL Server in question then you might find another computer running SQL Server on which you can install the...

But your formula returns the Sunday prior to the previous Sunday. And be careful when using DateDiff(week... - it's a 'wicked' creature.

Try this formula:

select dateadd(d,-1*datediff(d,'20080106',getdate())%7,getdate())

What it does is

- take difference...

isql.exe (SS2000) or sqlcmd.exe (SS2005) can be used to do this;

type them in with /? as a parameter which will show you all the options;

where lower(replace(name,'''',''))='martins'

On another thread I posted my X-paradigm for easily remembering how to code if two dates overlap. Thus two intervals S1-F1 and S2-F2 overlap if S1 < F2 and S2...