What I want to do is find all the groups where the inserts all
happened with in say 10 seconds. So my group by would be more like......
Perhaps the easiest solution is to make a temp table of datetime ranges
from the resultset, then join from and group by those rowIDs.
Chris W wrote:
> I have the following query...
> SELECT CreateDate, count( * )
> FROM `userprofile`
> GROUP BY CreateDate
> It isn't exactly what I want. Records are added to this table in 2 main
> ways. First people use the web site interface to create records. In
> this case, records are only added by one or 2 people and with a
> significant time between record inserts. The second way is through an
> import that reads data from a text file. In the second case the date on
> the records will all be close together with about 60 records added per
> second. What I want to do is find all the groups where the inserts all
> happened with in say 10 seconds. So my group by would be more
> GROUP BY CreateDate +- 10 SECOND
> Is there a way to do this?