Reputation: 45
I actually have a query joining 3 tables to collect informations so I can calculate some KPIs, but my logic is flawed, here is my actual query :
SELECT t.idCustomer, t.nameCustomer
COUNT(DISTINCT t.idTrip),
SUM(
CASE
WHEN t.tripDone <> 1
THEN 1
ELSE 0
END),
SUM(CASE
WHEN t.codeIncident = 'CANCEL'
THEN 1
ELSE 0
END)
FROM
(SELECT customer.idCustomer, customer.nameCustomer, trip.tripDone, incident.codeIncident
FROM CUSTOMER customer
JOIN TRIP trip ON customer.idCustomer = trip.idCustomer
JOIN INCIDENT incident ON trip.idTrip = incident.idTrip) t
GROUP BY t.idCustomer, t.nameCustomer
So, I want to know for each Customer :
The big mistake I made here, is that a trip can have multiple codeIncidents (example : one record for an idTrip with the codeIncident 'CANCEL' and another record with same idTrip with the codeIncident 'DELAYED'), so when I calculate the Sum when t.tripDone <> 1 I get a result of : '2' instead of '1' (because there are 2 records in my from Clause that have the t.tripDone <> 1 for the same idTrip).
Would you have any idea on how I should process this query so I can do the Sum when tripDone <> 1 only once for each tripId ?
Thanks a lot for the help !
If you need some more infos I'm available, and sorry for my lack of english skills !
Upvotes: 2
Views: 89
Reputation: 1
First Select idTrip field in your inner query that means table "t"
Upvotes: 0
Reputation: 191235
It sounds like you want to do the same count(distinct ...)
pattern for the columns you're currently summing, but with some logic. You can use case
within a count instead in the same way:
...
COUNT(
DISTINCT CASE
WHEN t.tripDone <> 1
THEN t.idTrip
ELSE null
END),
COUNT(
DISTINCT CASE
WHEN t.codeIncident = 'CANCEL'
THEN t.idTrip
ELSE null
END)
The else null
is a bit redundant as that's the default. As count()
ignores nulls, if the when
isn't matched then that trip ID isn't counted.
Upvotes: 1