没有JOIN的SQL查询

Question

I've a table:-

id, event_type

0, sent
1, sent
2, sent
3, sent
4, sent
2, delivered
4, delivered

Now, I want all those ids in which there was a sent event_type, but not delivered , ie 0, 1, 3,

I can do it with a nested or a MINUS query:-

select id from table where event_type = 'sent'
MINUS
select id from table where event_type = 'delivered';

But these are 2 different queries. How can it be done in just 1 single query?

I know there's a possibility of a JOIN query. Is there also a possibility without using a JOIN query and still getting the requisite output?

Answer 1

You can do conditional aggregation based filtering using HAVING clause with GROUP BY :

SELECT id 
FROM your_table_name 
/* consider only those id(s) which has either sent and/or delivered */
WHERE event_type IN ('sent', 'delivered')
GROUP BY id 
HAVING SUM(event_type = 'sent')  /*Consider id which has 'sent' */
       AND NOT SUM(event_type = 'delivered') /*Ignore id which has 'delivered' */

---

Another Option is to use a Correlated Subquery with NOT EXISTS and DISTINCT :

SELECT DISTINCT t1.id 
FROM your_table_name AS t1
WHERE t1.event_type = 'sent' 
  AND NOT EXISTS (SELECT 1 FROM your_table_name AS t2 
                  WHERE t2.id = t1.id AND 
                        t2.event_type = 'delivered')

Answer 2

You should have done it via group by as you dont have multiple tables to refer columns and get values together neither you want to manipute column using self join to get another instance of your same table rather your requirement is to fetch resulTs of specific groups with some condition.

  SELECT id, max(event_type) from table 
   group 
    by id 
   having count(distinct event_type)=1 
  and max(distinct event_type)='sent'