Query with/without group by returning totally different results

Question

Hi,
I've got this query:
  select count(DISTINCT UL.user_id)
  from app_user_log UL,
    company_registry CR,
    productarea PA,
    user_registry UR
       where UL.time_stamp >= to_date('04/02/2002','dd-mm-yy') 
       and UL.time_stamp <= to_date('10/02/2002','dd-mm-yy') 
     and UL.description = 'Open'
     and UR.INT_COMPANY_ID=CR.COMPANY_ID
     and PA.PRODUCTAREA_ID=CR.PRODUCTAREA_ID
     and UL.USER_ID=UR.USER_IDIt returns the value 262
I then want to divide this into one block pr application_id so that I get the number of users with app_id=1 as first applications, the number of users with app_id=2 as first application, and so on..
I thought this would be it:
  select ul.application_id, count(DISTINCT UL.user_id)
  from app_user_log UL,
    company_registry CR,
    productarea PA,
    user_registry UR
       where UL.time_stamp >= to_date('04/02/2002','dd-mm-yy') 
       and UL.time_stamp <= to_date('10/02/2002','dd-mm-yy') 
     and UL.description = 'Open'
     and UR.INT_COMPANY_ID=CR.COMPANY_ID
     and PA.PRODUCTAREA_ID=CR.PRODUCTAREA_ID
     and UL.USER_ID=UR.USER_ID
      group by ul.application_idBut this query yields an output with the sum of the counts way above 262 as I would expect..

Answer 1

Approach the problem as you would any programming problem - break it down into simpler problems. Solve each of these logically.
Simplifying here, but you should get the basics of this. For example. We need the earliest time stamp from a table. Then we need to select a distinct list of users and applications for that time stamp.
with EARLIEST_TIME_STAMP as (
select
MIN( datetime_stamp ) AS datetime
from my_table
select
count( DISTINCT userid ) as DISTINCT_USERS,
count( DISTICT applications ) as DISTINCT_APPLICACTION
from my_table t
join EARLIEST_TIME_STAMP s
on s.datetime = t.datetime_stamp