In the meantime, checkout DATEPART in books online. That and GROUP BY will be in your solution I think.
Can't we use normal conventional Group by here ?
I see no mention of using an unconventional GROUP BY, so I don't understand the reason for your comment.
If the underlying date data types contain time information – and we must assume that they do, based on the requirement – then DATEPART can be used to break these datetimes out into hourly blocks, before grouping them.
It was actually a wrong ping / comment to your quote, Apologies. thought of replying OP:hehe: