The problem here is that you are trying to update a table you aren't even referencing in the UPDATE statement's FROM clause.   I'm not really sure why you're asking the...

Not sure about whether performance will matter here, but the following code might be better on performance:

WITH ORDERED_DATA AS (

SELECT
T.Company,
T.[Month],
T.Sales,
ROW_NUMBER() OVER(PARTITION BY T.[Month] ORDER BY T.Sales DESC)...

This is a most interesting topic, and I just have to toss a hat in the ring.   Decided to try to make the code as simple as possible and see...

• This reply was modified 5 years, 10 months ago by sgmunson. Reason: Retry sqlplan file upload
• This reply was modified 5 years, 10 months ago by sgmunson. Reason: Had to use the XML format for a code window to get the sql plan file in place

Pietlinden,

Joining a second time isn't going to get another row...  not that we know what the OP needs here, but just sayin' ...

Try the following:

CREATE TABLE dbo.emp_availability (
emp_no INT,
    day_of_week INT,
    from_time VARCHAR(20),
    to_time VARCHAR(20)
);
INSERT INTO dbo.emp_availability (emp_no, day_of_week, from_time,...

Also, you might be able to use scalar functions you create in tempdb, based on making your parsing logic smart enough to detect a query length long enough to cause...

Outside of CROSS APPLY, you'd probably have to use Global Temp Tables - they're that way when the table name starts with "##" (2 pound signs).   Meaning that you run...

Without further detail, I'm not sure what you mean when you say "How can i get the query to show two different results."   Does that mean separate rows?   As you've...

Not in a tablix...   You MUST use a matrix to do this.    Grouping by Year as a "row group", and then your other groups would need to be "column groups".   ...

As I said before, you need a matrix and not a tablix.   However, you can effectively turn the tablix into a matrix by adding a column group.   The year would...

Once again, I'll agree with Brian.    Seems likely that a trigger on the table being updated might be doing something that tracks changes, and possibly to a varchar(max) or nvarchar(max)...

I agree with Brian.   "Pluggable into any SQL database" is just not going to realistically be possible when you're looking for an MS SQL specific feature AND you're going to...

Okay, so we're talking about 2,000,000 rows of 55 columns of up to 100 characters each, plus all the element overhead.   This could easily be a full Gigabyte of RAM...

Okay, to start, you're not grouping by year because you are putting the year in the same row as the other groups...  you probably need a matrix instead of a...

Okay, so let me get this straight.   Your data is insufficiently unique, so rather than actually fixing the real problem, you'll go to the transaction log and DBCC page commands...

• This reply was modified 5 years, 10 months ago by sgmunson. Reason: Didn't see Scott responding until I posted