Making changes on a pivot table - oracle

Question

I have the following query:
SELECT CASE GROUPING_ID(status, CASE WHEN status IN (2, 4, 5) THEN 1 ELSE 0 END)
WHEN 0
THEN TO_CHAR(status)
WHEN 2
THEN 'subtotal'
ELSE 'total'
END AS status,
COUNT(CASE user_type WHEN 1 THEN 1 END) AS "1",
COUNT(CASE user_type WHEN 2 THEN 1 END) AS "2",
COUNT(CASE user_type WHEN 3 THEN 1 END) AS "3",
COUNT(CASE user_type WHEN 5 THEN 1 END) AS "5",
COUNT(*) AS total
FROM (Select STATUS,USER_TYPE
FROM TRANSACTIONS tr join TRANSACTION_STATUS_CODES sc on sc.id = tr.user_type join
TRANSACTION_USER_TYPES ut on ut.id=tr.user_type
WHERE tr.status!=1 AND Tr.Update_Date BETWEEN TO_DATE('2022-01-01 00:00:00', 'yyyy-mm-dd HH24:MI:SS') AND TO_DATE('2022-11-13 23:59:59', 'yyyy-mm-dd HH24:MI:SS'))
GROUP BY CUBE(status,CASE WHEN status IN (2, 4, 5) THEN 1 ELSE 0 END)
HAVING GROUPING_ID(status, CASE WHEN status IN (2, 4, 5) THEN 1 ELSE 0 END) IN (0, 3)
OR ( GROUPING_ID(status, CASE WHEN status IN (2, 4, 5) THEN 1 ELSE 0 END) = 2
AND CASE WHEN status IN (2, 4, 5) THEN 1 ELSE 0 END = 1 );
which displays:
[
with inner query giving ( thats 10 out of 36rows):
anyways.. I've been asked to change the inner query to a better organized one, so I came up with the following pivot table ( which I want it to be my new inner query ):
SELECT p.*
FROM (
SELECT user_type,
status
FROM transactions
WHERE status !=1
AND Update_Date >= DATE '2022-01-01'
AND Update_Date < DATE '2022-11-14'
) PIVOT (
COUNT(*)
FOR user_type IN (1,2,3,5)
) p
ORDER BY status asc;
which gives:
what I need to try and achieve is the sum of subtotal/total of rows aswell as total of column( as I have in the first table ), and when I try to make it an inner query the outer query doesnt recognize anything..
so basically to add these:
is that achievable without changing the entire query?
or else, what is the best approach to achieve that?
thanks in advance!

Answer 1

CUBE does what you want in a single query and was covered in my previous answer.
You cannot use PIVOT to do what you want in a single query; instead you would need to use multiple queries joined together with UNION ALL to calculate the status, sub-total and total parts (which is likely to be much less efficient):
SELECT TO_CHAR(status) AS status,
"1",
"2",
"3",
"5",
"1" + "2" + "3" + "5" AS total
FROM table_name
PIVOT ( COUNT(*) FOR user_type IN (1,2,3,5) )
UNION ALL
SELECT 'SUB_TOTAL',
"1",
"2",
"3",
"5",
"1" + "2" + "3" + "5" AS total
FROM (SELECT user_type FROM table_name WHERE status IN (2,4,5))
PIVOT ( COUNT(*) FOR user_type IN (1,2,3,5) )
UNION ALL
SELECT 'TOTAL',
"1",
"2",
"3",
"5",
"1" + "2" + "3" + "5" AS total
FROM (SELECT user_type FROM table_name)
PIVOT ( COUNT(*) FOR user_type IN (1,2,3,5) )
or:
SELECT TO_CHAR(status) AS status,
COUNT(CASE user_type WHEN 1 THEN 1 END) AS "1",
COUNT(CASE user_type WHEN 2 THEN 1 END) AS "2",
COUNT(CASE user_type WHEN 3 THEN 1 END) AS "3",
COUNT(CASE user_type WHEN 5 THEN 1 END) AS "5",
COUNT(*) AS total
FROM table_name
GROUP BY status
UNION ALL
SELECT 'SUB_TOTAL',
COUNT(CASE user_type WHEN 1 THEN 1 END) AS "1",
COUNT(CASE user_type WHEN 2 THEN 1 END) AS "2",
COUNT(CASE user_type WHEN 3 THEN 1 END) AS "3",
COUNT(CASE user_type WHEN 5 THEN 1 END) AS "5",
COUNT(*) AS total
FROM table_name
WHERE status IN (2,4,5)
UNION ALL
SELECT 'TOTAL',
COUNT(CASE user_type WHEN 1 THEN 1 END) AS "1",
COUNT(CASE user_type WHEN 2 THEN 1 END) AS "2",
COUNT(CASE user_type WHEN 3 THEN 1 END) AS "3",
COUNT(CASE user_type WHEN 5 THEN 1 END) AS "5",
COUNT(*) AS total
FROM table_name
Which, for the sample data, (also from my previous answer):
CREATE TABLE table_name (status, user_type) AS
SELECT 2, 1 FROM DUAL CONNECT BY LEVEL <= 3 UNION ALL
SELECT 4, 1 FROM DUAL CONNECT BY LEVEL <= 13 UNION ALL
SELECT 5, 1 FROM DUAL CONNECT BY LEVEL <= 1 UNION ALL
SELECT 3, 1 FROM DUAL CONNECT BY LEVEL <= 5 UNION ALL
SELECT 3, 5 FROM DUAL CONNECT BY LEVEL <= 1 UNION ALL
SELECT 0, 1 FROM DUAL CONNECT BY LEVEL <= 4 UNION ALL
SELECT 0, 5 FROM DUAL CONNECT BY LEVEL <= 8;
Both output the same as using a single query with CUBE:
STATUS
1
2
3
5
TOTAL
2
3
0
0
0
3
4
13
0
0
0
13
0
4
0
0
8
12
5
1
0
0
0
1
3
5
0
0
1
6
SUB_TOTAL
17
0
0
0
17
TOTAL
26
0
0
9
35
If you did want to do it without CUBE then you can generate extra rows before pivoting (however, I think CUBE will still be more efficient):
SELECT title AS status,
"1",
"2",
"3",
"5",
"1" + "2" + "3" + "5" AS total
FROM (
SELECT l.*, t.user_type
FROM table_name t
CROSS JOIN LATERAL (
SELECT t.status, TO_CHAR(t.status) AS title, 1 AS priority FROM DUAL UNION ALL
SELECT NULL, 'SUB_TOTAL', 2 FROM DUAL WHERE t.status IN (2,4,5) UNION ALL
SELECT NULL, 'TOTAL', 3 FROM DUAL
) l
)
PIVOT ( COUNT(*) FOR user_type IN (1,2,3,5) )
ORDER BY priority, status
Which produces the same output.
fiddle