Question

I would like to formulate a query that returns the vin, customer_id and model_id where the same customer_id has multiple records that have a model_id of “toyota” from the CAR table. The vin in the primary key and unique. Any ideas? I’m using Oracle 10g.

CAR Table:

vin varchar (primary key)
customer_id varchar
model_id varchar

Data:

vin        customer_id   model_id 
------------------------------------  
222313XYZ,   cust1234,   toyota
123232SSS,   cust1234,   toyota
111111ERE,   cust1111,   bmw
999999LKO,   cust1224,   bmw

Answer 1

SQL Fiddle

Oracle 11g R2 Schema Setup:

CREATE TABLE car
    (vin varchar2(9), customer_id varchar2(8), model_id varchar2(6))
;

INSERT ALL 
    INTO car (vin, customer_id, model_id)
         VALUES ('222313XYZ', 'cust1234', 'toyota')
    INTO car (vin, customer_id, model_id)
         VALUES ('123232SSS', 'cust1234', 'toyota')
    INTO car (vin, customer_id, model_id)
         VALUES ('111111ERE', 'cust1111', 'bmw')
    INTO car (vin, customer_id, model_id)
         VALUES ('999999LKO', 'cust1234', 'bmw')
SELECT * FROM dual
;

Query 1: in case you want to see all models whose owners also own more than one toyota

select vin, customer_id, model_id from (
 select vin, customer_id, model_id, 
 count( decode(model_id, 'toyota', 1) ) over (partition by customer_id) cnt
  from car
)
where cnt > 1

Results:

|       VIN | CUSTOMER_ID | MODEL_ID |
--------------------------------------
| 999999LKO |    cust1234 |      bmw |
| 222313XYZ |    cust1234 |   toyota |
| 123232SSS |    cust1234 |   toyota |

Query 2: in case you want to see only toyotas whose owners own more than one of them:

select vin, customer_id, model_id from (
 select vin, customer_id, model_id, 
 count(*) over (partition by customer_id) cnt
  from car
 where model_id = 'toyota'
)
where cnt > 1

Results:

|       VIN | CUSTOMER_ID | MODEL_ID |
--------------------------------------
| 123232SSS |    cust1234 |   toyota |
| 222313XYZ |    cust1234 |   toyota |