一个SQL书写的问题

Question

有AB两个表结构如下

CREATE TABLE a
(
    id varchar(1) NULL
  , iyear varchar(4) NULL
  , imonth varchar(2) NULL
);
CREATE TABLE b
(
    id varchar(1) NULL
  , iyear varchar(4) NULL
  , imonth varchar(2) NULL
  , num int4 NULL
);

数据内容如下

INSERT INTO b (id, iyear, imonth, num) VALUES('1', '2019', '08', 5);
INSERT INTO b (id, iyear, imonth, num) VALUES('1', '2020', '05', 4);
INSERT INTO b (id, iyear, imonth, num) VALUES('1', '2020', '11', 3);
INSERT INTO b (id, iyear, imonth, num) VALUES('1', '2021', '09', 2);


INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '01');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '02');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '03');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '04');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '05');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '06');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '07');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '08');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '09');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '10');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '11');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2019', '12');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '01');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '02');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '03');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '04');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '05');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '06');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '07');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '08');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '09');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '10');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '11');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2020', '12');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '01');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '02');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '03');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '04');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '05');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '06');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '07');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '08');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '09');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '10');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '11');
INSERT INTO a (id, iyear, imonth) VALUES('1', '2021', '12');

希望实现如下效果

id	iyear	imonth	iyear_b	imonth_b	num	row_nu
1	2019	01	2019	08	5	1
1	2019	02	2019	08	5	1
1	2019	03	2019	08	5	1
1	2019	04	2019	08	5	1
1	2019	05	2019	08	5	1
1	2019	06	2019	08	5	1
1	2019	07	2019	08	5	1
1	2019	08	2019	08	5	1
1	2019	09	2020	05	4	1
1	2019	10	2020	05	4	1
1	2019	11	2020	05	4	1
1	2019	12	2020	05	4	1
1	2020	01	2020	05	4	1
1	2020	02	2020	05	4	1
1	2020	03	2020	05	4	1
1	2020	04	2020	05	4	1
1	2020	05	2020	05	4	1
1	2020	06	2020	11	3	1
1	2020	07	2020	11	3	1
1	2020	08	2020	11	3	1
1	2020	09	2020	11	3	1
1	2020	10	2020	11	3	1
1	2020	11	2020	11	3	1
1	2020	12	2021	09	2	1
1	2021	01	2021	09	2	1
1	2021	02	2021	09	2	1
1	2021	03	2021	09	2	1
1	2021	04	2021	09	2	1
1	2021	05	2021	09	2	1
1	2021	06	2021	09	2	1
1	2021	07	2021	09	2	1
1	2021	08	2021	09	2	1
1	2021	09	2021	09	2	1
1	2021	10				1
1	2021	11				1
1	2021	12				1

即用发生月份时的数值来填充之前的月份，直到遇到不为空的月份，比如19年8月有发生，则19年1-7月均为8月的值，20年5月有发生则19年9月-20年4月均为20年5月的值

SELECT
    *
FROM
(
    SELECT
        a.id
      , a.iyear
      , a.imonth
      , b.iyear iyear_b
      , b.imonth imonth_b
      , b.num
      , ROW_NUMBER() OVER(PARTITION BY a.id,a.iyear,a.imonth order by concat(b.iyear,b.imonth) ) row_nu
    FROM  a
    LEFT JOIN  b
      ON  a.id = b.id
     AND  concat(a.iyear,a.imonth) <= concat(b.iyear,b.imonth)
) A
WHERE row_nu = 1
;

我写了一个解决方法，可以通过窗口函数排序后得到正确的值,但当B表中数据多的时候会出现多条关联结果A部分的数据就会翻倍许多，是否有更好的写法来实现相同的效果呢？