Constituents by order date craze

Dearest Friends, Colleagues, and Tessiturans!

So I'm in the middle of a data collection that FEELS like it should be easy, but I'm not exactly getting the results I anticipated so I wanted to crowd source my findings and confusions to see if I'm crazy or not, and if I'm redeemable either way!

The impetus of my exploration is matching (or not match!) some social media engagement with purchase history, and so I'm basically pulling anybody who purchased in our recent season between January 1 and June 25 (to correspond with the timing for various social media campaigns). I've done this in SQL with the following query:

select * from t_ORDER
join t_ticket_history
on t_order.customer_no = t_ticket_history.customer_no
where t_ticket_history.season = 55
and
t_order.order_dt between '2017-01-01 00:00:00.000' and '2017-06-25 23:00:00.000'

This gives me a comprehensive list of all orders for the desired timeframe for my previous season, hooray! BUT--when I check through the list to verify things are correct, I'm noticing some sub exchanges showing up for single performances with the corresponding order total that was renewed prior to the period I'm interested in.

For example, I have a subscriber near the top of my list that renewed back in August of 2016 for a total of $768. Outside of the timeframe I'm interested in. When May of 2017 came along, however, he decided to exchange out of one performance and into another, as he is often want to do with scheduling conflicts he experiences (not social media engaging him). 

If I look at my spreadsheet, it shows his order date in May for the exchange, and shows a total ticket purchase of $768. This might appear to the untrained eye as a purchase in May for $768, when actually it was an exchange for $0. Because I have subs exchanging in and out throughout the season in a similar fashion, I'd love to pair down my results to eliminate any such false positives, so to speak.

Any thoughts, feelings, reflections, questions, comments, or otherwise?

THANK YOU!

Brian Jones