Don't get into the habit of using implicit casting. If you're joining on two columns with different data types, one entire column has to be cast, and if you have...

It's not that CAST doesn't work, it's that it's expensive, especially if you use it in a crazy place like a join.

It's easier than that.

select * from tblTest
where TestNumber = '20012';

You're welcome... sorry it took me a bit to actually understand the question.

(sorry, my internet is freaking out... double post).

• This reply was modified 6 years, 2 months ago by pietlinden. Reason: duplicate post

I think this works... <g> Standard pattern is to do the TOP N inside the CROSS APPLY, and join the Cross Apply to the outer query. (that's what the tt.Month...

• This reply was modified 6 years, 2 months ago by pietlinden.

This looks like it works:

SELECT DISTINCT t.Company,sales.[Month], sales.Sales
FROM TestTable t
CROSS APPLY (SELECT TOP 5 *
FROM [TestTable] t2
WHERE t.Company = t2.Company
ORDER BY t2.Sales DESC) sales
ORDER BY t.Company, sales.Sales...

Oh, I see now... how about something like this:

CREATE TABLE Employee (
EmployeeKey INT IDENTITY PRIMARY KEY,
FirstNameVARCHAR(20) NOT NULL,
LastNameVARCHAR(20) NOT NULL,
ShiftNo TINYINT NOT NULL
);

-- both WelderID and OperatorID...

Well, I meant joining to that table twice and returning the two different employee names...

TB_EMP.EMP_INT AS "WELDER",

TB_EMP.EMP_INT AS "OPERATOR",

Oh wait... can't do that... because that's only one column in EMP...

Store the dates as datetime, not VARCHAR. Then you can just use DATEDIFF(minute,startdate, enddate)

Otherwise you have to cast the VARCHAR columns, and that'll cause the query to be really slow...

Looks like you need two instances of TB_EMP to join on the welder key and another to join on the Operator key.

1 lakh = 100,000

Are you trying to create an aging query, so that each invoice is "binned" by age, which is determined by [DocDate]? (the usual is something like "less than 30 days",...

• This reply was modified 6 years, 2 months ago by pietlinden. Reason: code reformat
• This reply was modified 6 years, 2 months ago by pietlinden.

Just to see if I could do it, I wrote some code to do the testing for me... <g>

I shamelessly stole Phil's checking code.... Seems to work, though.

            
July 22, 2020 at 6:34 pm
            
#3773927

            
            
                                                
        

Steve,

Since it's only looking at column names and not rows of data, it seems this could be done easily with a little bit of dynamic SQL (just get a list...