Does the web user have the same priveleges as the user that QA is running under?

If so, and If the sp works OK in QA but not through a web...

A very interesting idea from rockmoose that I never thought of doing. However, I can see over time that the cube will become very hard to investigate for differences.

A good...

Does a check in the maintenance job every night and comes back with 0 errors.

Apologies. PLease ignore my comment. I was thinking of something else I have been doing lately.

A better option would be to use the 'REFRESH' method. This will take longer than...

There is no way to tell AS how to do this for an incremental update.

What you need to do is find all the records that were changed and...

Either:

Make sure that you 'set identity_insert off' before the append

Or:

If you are writing insert into xxx (id,col1, col2).... do NOT include the id column in the list of fields...

I usually disable the triggers and build the sql the triggers are performing into the insert statement - You will also find that it runs quicker. I have tried the...

If you want to avoid cursors, you can try using a data driven query in dts. See link:

http://doc.ddart.net/mssql/sql2000/html/dtssql/dts_elemtsk1_9w2z.htm

This enables you to check for existience of a record and run an...

Another way that gives you a choice on which column you want the top to be:

/* create the procedure to return top rows for a given number

hint--column_name and table_name to...

You can only have one primary key on a table.

You need to check BOL for UNIQUE columns. There are probabaly what you are after.

Have you thought about creating a table to act as a lookup. That way if your insert fails to find a record, you can reference the lookup table. Alternatively, you...

You can download a trial version of visio and try that.

Alternatively, I have in the past used linked tables to sql from ms access and brought the tables up in...

A cursor would appear to be the only way. A bit slower but gets the job done OK.

/*

source table creation

create table groupings

(id int, col1(varchar(1), col2(varchar(1), col3(varchar(1))

insert into groupings values (1,...

It's hard to know exactly what you are trying to achieve and how much data we are talking about.

In any case, this should work for reasonably sized tables.

1. Create a...

Sounds very similar to the problem I had. See link to this site

http://www.sqlservercentral.com/forums/shwmessage.aspx?forumid=19&messageid=120028#bm120824

It looks like a permisssions problem. When you run the package from designer it uses the...