Viewing 15 posts - 2,731 through 2,745 (of 7,609 total)
To avoid any IF / CASE logic, I usually use this technique instead:
SIGN(Jan) + SIGN(Feb) + ...
In situations where the numbers could ever be negative, you'd need to include ABS()...
August 19, 2019 at 5:21 pm
I would think so, since SQL doesn't (re)order / sort rows going into a columnstore, it loads them in the same order in which they arrive.
Thus, presumably the first 1,048,576...
August 16, 2019 at 8:09 pm
Yeah, I was incorrectly thinking of column3 also matching.
I don't like having to have duplicative indexes. I'd probably go with just:
( column2, datetime ) INCLUDE ( column3 )
Sure, that index...
August 16, 2019 at 7:18 pm
I would of that the best index would be:
(Column3, Column2, DateTime)
given the WHERE condition.
Particularly if you later use the same query with:
Where tab1.Column3= 9
to query a different column.
And I'd move...
August 16, 2019 at 5:38 pm
Since an IN gets converted to ... OR ... OR ..., I suppose it might make a difference, in which case you'd want to put the most common values first.
You...
August 15, 2019 at 8:22 pm
Technically, if j.Grp is NULLable, I think you'd need to use:
WHERE j.Grp IS NOT NULL AND j.Code <> 'L3'
to insure the same results as the original code.
August 15, 2019 at 8:13 pm
You waste space and risk ambiguity by storing dashes in a date? Seriously? If for some bizarre reason one insists on storing dates as char, they should be YYYYMMDD. Unambiguous. ...
August 6, 2019 at 9:29 pm
I don't really have any issue with storing cc numbers as char(16), especially since they'll need to be encrypted and end up being stored as binary anyway. I also wouldn't...
August 6, 2019 at 6:35 pm
Actually an identifier is numeric (integer) because it just makes no common sense to do otherwise. Overly-pedantic concerns about whether it's used in math or not are actually irrelevant. There...
August 6, 2019 at 4:54 pm
Probably cleaner to check for 0 values as you INSERT the row:
INSERT INTo #a
select NULLIF(0,0),NULLIF(1,0),NULLIF(0,0),NULLIF(2,0)
union ALL
select NULLIF(0,0),NULLIF(2,0),NULLIF(0,0),NULLIF(3,0)
union ALL
select NULLIF(0,0),NULLIF(3,0),NULLIF(0,0),NULLIF(4,0)
August 6, 2019 at 1:57 pm
A trigger is actually perfect for this situation. For the email, it's probably best to add row(s) to a table or queue, then construct and send the emails based on...
August 1, 2019 at 7:43 pm
It seems like it, but SQL 2008 doesn't directly read JSON, afaik. It might be easier to handle it with T-SQL rather than try to create some external function/process.
August 1, 2019 at 7:18 pm
yyyy-mm-00 is not an interval either. The interval would be yyyy-mm-01 to yyyy-mm-{EOM}.
yyyy-mm-00 sorta looks like a date, but it is not one, no matter how much fluff you put...
August 1, 2019 at 5:55 pm
Limited sample data, but I think this should perform better, since it avoids all the CHARINDEX costs:
SELECT A.ID, B.ItemNumber,
MAX(CASE WHEN D.Item...
August 1, 2019 at 5:39 pm
The disadvantage is that it violates 1NF big time. A "date" such as yyyy-mm-00 is not a date it is multiple dates. Thus, it violates 1NF.
Besides which, SQL Server does...
July 31, 2019 at 8:49 pm
Viewing 15 posts - 2,731 through 2,745 (of 7,609 total)