De-duping customers based on a second field

Question

Not sure what to call this, but say I have a table that has two fields: Customer ID and Movie Purchased (so customers are not unique in the table):
CUST ID ....... MOVIE
====== ....... =====
1 ....... Dark Knight
1 ....... Swing Vote
2 ....... Pineapple Express
2 ....... Swing Vote
3 ....... The Mummy
3 ....... Mamma Mia
etc.
I'd like to write a query that returns a distinct customer ID and the movie purchased based on a hierarchy. So, if the hierarchy goes 1-Dark Knight, 2-Swing Vote, 3-Pineapple Express, I want the customer ID and that movie name if they purchased it. In this example I'd like to return :
CUST ID ....... MOVIE
====== ....... =====
1 ....... Dark Knight
2 ....... Swing Vote
(no cust ID 3 because their movies weren't in the hierarchy)
Any hints as to the best way to do this? I originally thought a CASE statement might somehow only retrieve the cust ID once but that does not seem to work. Any thoughts would be appreciated.
Thanks!

Answer 1

I think I'm interpreting what you wanted differently than isotope, heres my interpretation with an extra record to help illustrate the difference in my thought process.
Attempt 1 - Still Playing with Analytic Functions so not sure if this will break easily
with hierarchy as (
      select 1 as rank, 'Dark Knight' as movie from dual union all
      select 2, 'Swing Vote'        from dual union all
      select 3, 'Pineapple Express' from dual),
    table1 as (
      select 1 as cust_id, 'Dark Knight' as movie from dual union all
      select 1, 'Swing Vote' from dual union all
      select 2, 'Pineapple Express' from dual union all
      select 2, 'Swing Vote' from dual union all
     select 3, 'The Mummy' from dual union all
     select 3, 'Mamma Mia' from dual union all
     select 4, 'Pineapple Express' from dual union all
     select 4, 'Dark Knight' from dual)
   select distinct t.cust_id, FIRST_VALUE( t.movie) OVER (PARTITION BY t.cust_id ORDER BY h.rank)
   from table1 t inner join hierarchy h on t.movie = h.movieAttempt 2
with hierarchy as (
      select 1 as rank, 'Dark Knight' as movie from dual union all
      select 2, 'Swing Vote'        from dual union all
      select 3, 'Pineapple Express' from dual),
    table1 as (
      select 1 as cust_id, 'Dark Knight' as movie from dual union all
      select 1, 'Swing Vote' from dual union all
      select 2, 'Pineapple Express' from dual union all
      select 2, 'Swing Vote' from dual union all
     select 3, 'The Mummy' from dual union all
     select 3, 'Mamma Mia' from dual union all
     select 4, 'Pineapple Express' from dual union all
     select 4, 'Dark Knight' from dual)
   select t.cust_id, t.movie
   from table1 t inner join hierarchy h on t.movie = h.movie
   where h.rank = (select min(rank) from table1 t1 inner join hierarchy h1 on t1.movie = h1.movie where t1.cust_id = t.cust_id)Both result in identical output of
1     Dark Knight
2     Swing Vote
4     Dark KnightI interpreted this slightly different than isotope in thinking that you were going to have an overall hierarchy not one per customer which I think is how isotope took it as he used intersect to match them up. Oh well either way maybe it will be useful to someone!
Rich
Message was edited by:
BengalGuy