Can record comparisons be done without functions or PLSQL

Question

I am used to writing Oracle functions to do record comparisons, however on this particular project the managers do not want to use functions or PLSQL
Is there a way to do the following with a SQL statement rather than writing a function - see data set below? Oracle 9i.
I have the following data set used to track packages that are moved from warehouse to warehouse on trucks
I only want to retrieve a record if the starting location is Dallas Ft Worth and the end location is another city.
I do not want to retrieve a record if the package moves between other cities (say Boston and New York)
For this data set package 1324 was added and removed to the warehouses at Dallas 4 times via truck 735 before it was finally moved to a warehouse in Boston by the same truck
I want the query to return the following record
105273     345          1324               725          11/3/2010 14:32               11/3/2010 15:46               Dallas Ft Worth          WH4
Showing the last location in Dallas before it was moved to Boston
I do not want to return any of the Boston or NY records or any of the previous Dallas Ft Worth Records.
The way the system updates the records is as follows:
If a package is added to a warehouse the record gets an added date.
When it is removed the system updates the removed date of that same record
Once it is added to a new warehouse there is a new record with an added date and null date for the removed date unless it leaves that location
In the record set if the removed date is null then that is the current location of the package
CREATE TABLE package_location
( track_id NUMBER(10)
, sender_id NUMBER(10)
, package_id NUMBER(10)
, truck_id NUMBER(10)
, package_add_date DATE,
, package_remove_date DATE,
, city VARCHAR2 (50)
, Warehouse VARCHAR2(50)
INSERT ALL
INTO package_location(track_id, sender_id, package_id, truck_id, package_add_date, package_remove_date, city, warehouse) VALUES (100012
, 345, 1324, 725, 8/4/2010 2:39:48 PM, 8/4/2010 2:53:07 PM, 'Dallas Ft Worth', 'WH1')
INTO package_location(track_id, sender_id, package_id, truck_id, package_add_date, package_remove_date, city, warehouse) VALUES (100016
, 345, 1324, 725, 8/4/2010 2:53:07 PM, 9/30/2010 10:41:01 AM, 'Dallas Ft Worth', 'WH2')
INTO package_location(track_id, sender_id, package_id, truck_id, package_add_date, package_remove_date, city, warehouse) VALUES (103055
, 345, 1324, 725, 9/30/2010 10:41:01 AM, 11/3/2010 2:32:14 PM, 'Dallas Ft Worth', 'WH3')
INTO package_location(track_id, sender_id, package_id, truck_id, package_add_date, package_remove_date, city, warehouse) VALUES (105273
, 345, 1324, 725, 11/3/2010 2:32:14 PM, 11/3/2010 3:46:22 PM, 'Dallas Ft Worth', 'WH4')
INTO package_location(track_id, sender_id, package_id, truck_id, package_add_date, package_remove_date, city, warehouse) VALUES (105284
, 345, 1324, 725, 11/3/2010 3:46:22 PM, 11/22/2010 11:55:18 AM, 'Boston', 'WH34')
INTO package_location(track_id, sender_id, package_id, truck_id, package_add_date, package_remove_date, city, warehouse) VALUES (106674
, 345, 1324, 725, 11/22/2010 11:55:18 AM, NULL, 'New York', 'WH56')
I tried to do a query with a subquery... the main query was getting all the records with Dallas Ft Worth and then the subquery was to check if the location was not Dallas Ft Worth, but so far I haven't quite been able to get the correct dataset.
Any help is greatly appreciated. I don't know if there is a way to do this with just a sql query, if not other suggestions are welcomed!
Thanks in advance

Answer 1

I added a couple of rows to your table, shipping the package back from New York to Dallas on a different truck, then back out from Dallas to New York just to make sure I understand the requirements.
SQL> SELECT * FROM package_location;
  TRACK_ID  SENDER_ID PACKAGE_ID   TRUCK_ID PACKAGE_ADD_DATE     PACKAGE_REMOVE_DATE  CITY             WAREH
    100012        345       1324        725 04-aug-2010 14:39:48 04-aug-2010 14:53:07 Dallas Ft Worth  WH1
    100016        345       1324        725 04-aug-2010 14:53:07 30-sep-2010 10:41:01 Dallas Ft Worth  WH2
    103055        345       1324        725 30-sep-2010 10:41:01 03-nov-2010 14:32:14 Dallas Ft Worth  WH3
    105273        345       1324        725 03-nov-2010 14:32:14 03-nov-2010 15:46:22 Dallas Ft Worth  WH4
    105284        345       1324        725 03-nov-2010 15:46:22 22-nov-2010 11:55:18 Boston           WH34
    106674        345       1324        725 22-nov-2010 11:55:18                      New York         WH56
    106675        345       1324        804 25-nov-2010 09:55:18 30-nov-2010 15:46:22 Dallas Ft Worth  WH4
    106676        345       1324        804 25-nov-2010 09:55:18 01-dec-2010 15:46:22 New York         WH42Given this data, you would want to get back track_id 105273 and 106675 correct?
If so, then something like:
SQL> SELECT track_id, sender_id, package_id, truck_id, package_add_date,
  2         package_remove_date, city, warehouse
  3  FROM (SELECT track_id, sender_id, package_id, truck_id, package_add_date,
  4               package_remove_date, city, warehouse,
  5               LEAD (city) OVER(PARTITION BY package_id
  6                                ORDER BY package_remove_date) next_city
  7        FROM package_location)
  8  WHERE city = 'Dallas Ft Worth' and
  9        next_city != city;
  TRACK_ID  SENDER_ID PACKAGE_ID   TRUCK_ID PACKAGE_ADD_DATE     PACKAGE_REMOVE_DATE  CITY             WAREH
    105273        345       1324        725 03-nov-2010 14:32:14 03-nov-2010 15:46:22 Dallas Ft Worth  WH4
    106675        345       1324        804 25-nov-2010 09:55:18 30-nov-2010 15:46:22 Dallas Ft Worth  WH4John
Edited by: John Spencer on Aug 14, 2011 2:22 PM
Fixed non-displaying operator