Question

MS SQL:

Table One: Routes –Table Two: Stops (Uses an int field [seq] to get the sequence of stops, key field ID)

How can I join [routes] to the [stops] table and return only the first and last route into a single row?

Let’s say I do a simple join:

Route 1 | Stop 1
Route 1 | Stop 2
Route 1 | Stop 3
Route 2 | Stop 1
Route 2 | Stop 2

I want it to return:

Route 1 | Stop 1 | Stop 3
Route 2 | Stop 1 | Stop 2

I can join them and get the Min(Seq) and Max(Seq), but not sure how I can get the ID and not Seq. It’s almost like I want to do two joins, but multiple fields. I need it to be efficient because of returning thousands of routes at a given time. So I don’t think a table valued function to get the stops would be beneficial. However, if I could get the min and max seq, returning the associated stop ID, I could simply join Stops to these to get the related data.

Answer 1

;WITH S
     AS (SELECT *,
                ROW_NUMBER() OVER (PARTITION BY RouteId ORDER BY seq) AS RN1,
                ROW_NUMBER() OVER (PARTITION BY RouteId ORDER BY seq DESC) AS RN2
         FROM   Stops)
SELECT R.Route,
       S1.Stop,
       S2.Stop
FROM   Routes R
       JOIN S S1
         ON R.Id = S1.RouteId
       JOIN S S2
         ON R.Id = S2.RouteId
WHERE  S1.RN1 = 1
       AND S2.RN2 = 1