[SQL][Bigquery] Case When to take preceding row for a condition

Question

So for a table like this i want to have another column which has the reqd_col value if cond_value is > 100 and to take the previous value in other cases.

User	cond_Value	Reqd_Col
123456	159	12:30
123456	34	12:32
123456	46	12:35
123456	98	12:37
123456	123	12:56
123456	12	13:00
789000	100	16:00
789000	54	16:10
789000	23	16:14
789000	122	17:05
789000	98	17:08
789000	133	17:23

So the output should look like

User	cond_Value	Reqd_Col	Output
123456	159	12:30	12:30
123456	34	12:32	12:30
123456	46	12:35	12:30
123456	98	12:37	12:30
123456	123	12:56	12:56
123456	12	13:00	12:56
789000	100	16:00	16:00
789000	54	16:10	16:00
789000	23	16:14	16:00
789000	122	17:05	17:05
789000	98	17:08	17:05
789000	133	17:23	17:23

Ideally i would like to have this as a view in BigQuery.

Answer 1

Try below query.

SELECT *, LAST_VALUE(IF(cond_Value >= 100, Reqd_Col, NULL) IGNORE NULLS) OVER w AS Output
  FROM sample_table
WINDOW w AS (PARTITION BY User ORDER BY PARSE_TIME('%R', Reqd_Col))
 ORDER BY 1, 3;

+--------+------------+----------+--------+
|  User  | cond_Value | Reqd_Col | Output |
+--------+------------+----------+--------+
| 123456 |        159 | 12:30    | 12:30  |
| 123456 |         34 | 12:32    | 12:30  |
| 123456 |         46 | 12:35    | 12:30  |
| 123456 |         98 | 12:37    | 12:30  |
| 123456 |        123 | 12:56    | 12:56  |
| 123456 |         12 | 13:00    | 12:56  |
| 789000 |        100 | 16:00    | 16:00  |
| 789000 |         54 | 16:10    | 16:00  |
| 789000 |         23 | 16:14    | 16:00  |
| 789000 |        122 | 17:05    | 17:05  |
| 789000 |         98 | 17:08    | 17:05  |
| 789000 |        133 | 17:23    | 17:23  |
+--------+------------+----------+--------+