Question

We have a dataset that looks like this:

ID       eventType     date
--------------------------------
1        foo           2 March 2013
2        foo           3 March 2013
3        bar           3 March 2013
4        foo           5 March 2013
5        foo           6 March 2013
6        bar           7 March 2013
7        baz           8 March 2013

I can easily get the unique list of eventTypes from this list. However, how do I query the count of which eventTypes first appeared BETWEEN startDate and endDate. I want to be able to have a date range from 7 March 2013 – 10 March 2013 and get a count returned of 1 since baz value was a newly occurring eventType during the date range. On the other hand, a date range of 5 March 2013 – 7 March 2013 would return a count of 0 since no newly occurring eventTypes appeared in that range.

Answer 1

I would just look at using the MIN aggregate function to find the earliest occurrence of the event type before the end date of the query. Then I would see if any events had their earliest event after the start date of the range.

SELECT event_type, date_value
from (
  SELECT event_type, min(date_value) as date_value
  from your_table
  where date_value <= date '2013-03-10'
  group by event_type
)
where date_value >= date '2013-03-07'