A part of the problem is doing the rounding inside the sum.  I recommend that the rounding be done on the sum rather than inside it for both reasons...

Yes - you need to group the results.  A couple of notes:

1. DAY(observ_date) returns an INT
2. MONTH(observ_date) returns an INT
3. Checking for MONTH >= 10 for the next fiscal year...

• This reply was modified 6 years ago by Jeffrey Williams. Reason: Modified the start/end date variables to use datetime2

What seems to be the problem?

Since you are on 2016 and you don't have access to STRING_AGG, you can still use FOR XML.

 Select Distinct
       ...

Here is what I was thinking - not tested and may need some refinement in the build of the temp tables.  This all depends on how many rows can satisfy...

For the creation of the #Entity temp table.

I would move the join to Industry into the build of the #EntityIndustry table - and then move the join to #EntityIndustry into...

Is there more to this code than you have shown?  I see several temp tables that are created - and never used.  I also see an outer join to a...

I have no idea why Microsoft has kept that option in the product - auto shrink will cause massive fragmentation of the database every time it shrinks.  It...

Just to confirm, COALESCE won't provide an improvement in performance over CASE, although it will make the code a whole lot easier.  Behind the scenes, the optimizer converts a...

One other optimization that *may* be possible is to remove the CASE expression.  I have to assume that if a match is found in #entity that the columns from that...

I have no idea why Microsoft has kept that option in the product - auto shrink will cause massive fragmentation of the database every time it shrinks.  It is just...

After going down that rabbit hole, I stumbled on a way to at least start the service by bypassing the upgrade script.  Once I was able to start MSSQLSERVER...

You can avoid errors where there are missing periods by making sure you always have 4 periods in the field.  Here is one approach:

 Select *
 ...

I guess I am confused - if you are going to be using SSIS then why would you need to use OPENROWSET to load the data into a table?  SSIS...

You need to specify the admin account that will be used - one that has sysadmin access to that instance of SQL Server.