If there are no related records found, the above query will not count the record from master table. Therefore, LEFT JOIN is more appropriate.

Try:

SELECT GroupItem AS Item

,SUM(QUANTITY) AS QUANTITY

FROM (

...

Sorry, Sorry, Sorry!

I didn't mean converting int 20091201 into DateTime as 01 Jan 2009. Above would never work of cause. What I mean (for INT) is different.

Try this:

...

You are absolutely right! You have right concerns! There are!

But I...

I wanted to create a bit more efficient query but the following made me to ask you one small question:

Chaitanya Patil (6/18/2010)

390 ...

How this would help for working with milliseconds? As soon as...

Absolutely agree.

However:

1. If you really need to use something to identify what happen before what, and that difference is down to millisecs and no other datetime functions are required...

You can store it as meaningfull value available for comparisons in "where" clause, but it cannot be datetime, so no datetime functions will be applicable.

What you do is: format it...

Well, well, well...

Actually, no surprise, BCP, old as Sybase (and also originated by them), is the fasterst way to get data out DB and back into it...

Cheers 🙂

Use bcp or SSIS

Now,

You have couple of choices:

1. Calculate time of single login and then sum it.

- It will illuminate the mismatches, but look quite stupid as you will have some thing...

1. To achieve the required order and format of data NO CURSOR is required in your case. As shown in my examples, you just introduce the artificial "order" column which...

Once upon a time there was a purple crocodile in a forest, he went to the office when the summer just started to open doors to the bus full of...

You can...

Just a small suggestion, if you will go with creating persisted computed column (and assuming your columns are in two diffrent tables), store the date as an INT using:

FLOOR(CAST(DateTimeValue...