How about storing them as positives as tinyint but interpret as negatives with computed columns, view, user stored procedure etc.?

If you accept to have differences in one row you could do something like:
SELECT T.*, S.*
FROM #Targer AS T
INNER JOIN #Source AS S
WHERE EXISTS...

What about solving a sudoku where there are no combos? Like this one?

Variable IS NOT converted to nchar. It is the other way arround. The literal is converted to char and thus the value set to variable is dependent on databse collation

Try...

Hey Kevin! This is really great stuff pointed there! Nice idea to take care of Sudoku in SQL.

I tried your solution on the harder-hardest-sudoku-ever and it seems to not...

@hugo

Thanks for the valuable comment. However, I didn't meant nor said that the behavior I observed is standard or repeatable. My goal was to just show how awkward...

g.britton (6/24/2015)

---

Nakul Vachhrajani (6/24/2015)

---

I guess the explanation is quite simple:

Divide by 0 is not causing a problem in the SELECT clause because of the presence of EXISTS. You can have...

Could you please explain the CBAC acronym used in this context? Thanks in advance!

Nice. But I would use if @binary like '%[^0^1]%' or @binary = '' or @binary is null

return -1

It will prevent from inputing any non-binary string

This book is great! I hope to win some.

--Example of some index statistics

select o.name, ColNo, InclColNo,

rank() over (order by ColNo desc) as RnkColNo,

rank() over (order by InclColNo desc) as...