Query populating dates - sql

Question

query that generates records to hold a future calculated value.
Hi I trying to write a query with the tables below to populate a collection. I want the t2 values when the dates match but when there is not a match I want the dates to populate with a null values (will be populate later with a calculated value) The number of records for the same date should match the last time the dates matched. So in the example for each day after 7/1 there should be 3 records for each day and after 7/5 just 2. I am trying to do this in one query but I am not sure it is possible. Any help on creating this and getting into a collection would be much appreciated.
create table t1 as
WITH DATA AS
(SELECT to_date('07/01/2019', 'MM/DD/YYYY') date1,
to_date('07/10/2019', 'MM/DD/YYYY') date2
FROM dual
)
SELECT date1+LEVEL-1 the_date,
TO_CHAR(date1+LEVEL-1, 'DY','NLS_DATE_LANGUAGE=AMERICAN') day
FROM DATA
WHERE TO_CHAR(date1+LEVEL-1, 'DY','NLS_DATE_LANGUAGE=AMERICAN')
NOT IN ('SAT', 'SUN')
CONNECT BY LEVEL <= date2-date1+1
create table t2
(cdate date,
camount number);
insert into t2 values
('01-JUL-2019', 10);
insert into t2 values
('01-JUL-2019', 20);
insert into t2 values
('01-JUL-2019', 30);
insert into t2 values
('05-JUL-19', 50);
insert into t2 values
('05-JUL-19', 20);
expected results:
01-JUL-19 10
01-JUL-19 20
01-JUL-19 30
02-JUL-19 null
02-JUL-19 null
02-JUL-19 null
03-JUL-19 null
03-JUL-19 null
03-JUL-19 null
04-JUL-19 null
04-JUL-19 null
04-JUL-19 null
05-JUL-19 50
05-JUL-19 20
08-JUL-19 null
08-JUL-19 null
09-JUL-19 null
09-JUL-19 null
10-JUL-19 null
10-JUL-19 null

Answer 1

One approach to this kind of problem is to build the result set incrementally in a few steps:
Count matches that each THE_DATE in T1 has in T2.
Apply the rule you outlined in the question to those THE_DATE which have zero matches (carry forward (across dates in ascending order) the number of matches of the last THE_DATE that did have matches.
Generate the extra rows in T1 for the THE_DATE that have zero matches. (e.g. If it is supposed to have three null records, duplicate up to this number)
Outer join to T2 to get the CAMOUNT where it is available.
Here's an example (The three named subfactors corresponding to steps 1,2,3 above):
WITH DATE_MATCH_COUNT AS (
SELECT T1.THE_DATE,
COUNT(T2.CDATE) AS MATCH_COUNT,
ROW_NUMBER() OVER (PARTITION BY NULL ORDER BY T1.THE_DATE ASC) AS ROWKEY
FROM T1
LEFT OUTER JOIN T2
ON T1.THE_DATE = T2.CDATE
GROUP BY T1.THE_DATE),
ADJUSTED_MATCH_COUNT AS (
SELECT THE_DATE,
MATCH_COUNT AS ACTUAL_MATCH_COUNT,
GREATEST(MATCH_COUNT,
(SELECT MAX(MATCH_COUNT) KEEP ( DENSE_RANK LAST ORDER BY ROWKEY ASC )
FROM DATE_MATCH_COUNT SCALAR_MATCH_COUNT
WHERE SCALAR_MATCH_COUNT.ROWKEY <= DATE_MATCH_COUNT.ROWKEY AND
SCALAR_MATCH_COUNT.MATCH_COUNT > 0)) AS FORCED_MATCH_COUNT
FROM DATE_MATCH_COUNT),
GENERATED_MATCH_ROW AS (
SELECT THE_DATE, FORCED_MATCH_COUNT, MATCH_KEY
FROM ADJUSTED_MATCH_COUNT CROSS APPLY (SELECT LEVEL AS MATCH_KEY
FROM DUAL CONNECT BY LEVEL <= DECODE(ACTUAL_MATCH_COUNT,0,FORCED_MATCH_COUNT,1)))
SELECT THE_DATE, CAMOUNT
FROM GENERATED_MATCH_ROW
LEFT OUTER JOIN T2
ON GENERATED_MATCH_ROW.THE_DATE = T2.CDATE
ORDER BY THE_DATE, CAMOUNT ASC;
Result:
THE_DATE CAMOUNT
____________ __________
01-JUL-19 10
01-JUL-19 20
01-JUL-19 30
02-JUL-19
02-JUL-19
02-JUL-19
03-JUL-19
03-JUL-19
03-JUL-19
04-JUL-19
04-JUL-19
04-JUL-19
05-JUL-19 20
05-JUL-19 50
08-JUL-19
08-JUL-19
09-JUL-19
09-JUL-19
10-JUL-19
10-JUL-19