SQL get first non value based on status - sql

Question

I'm not sure if the title is correct but here's my question. I have a table like this:
+----+--------+--------------+---------+------------+
| id | city | province | status | date |
+----+--------+--------------+---------+------------+
| 1 | cainta | rizal | failed | 01/01/2020 |
| 1 | null | null | success | 02/01/2020 |
| 1 | cainta | rizal | failed | 03/01/2020 |
| 2 | pasig | metro manila | failed | 04/01/2020 |
| 2 | pasig | metro manila | failed | 05/01/2020 |
| 2 | null | null | success | 06/01/2020 |
| 3 | obando | bulacan | failed | 07/01/2020 |
| 3 | null | null | failed | 08/01/2020 |
| 3 | obando | bulacan | success | 09/01/2020 |
+----+--------+--------------+---------+------------+
Now I need to get all transactions with status='success'. If I do that the output will be like this:
| id | city | province | status | date |
|------|--------|------------|----------|------------|
| 1 | nan | nan | success | 02/01/2020 |
| 2 | nan | nan | success | 06/01/2020 |
| 3 | obando | bulacan | success | 09/01/2020 |
What I need is this:
| id | city | province | status | date |
|------|--------|--------------|----------|------------|
| 1 | cainta | rizal | success | 02/01/2020 |
| 2 | pasig | metro manila | success | 06/01/2020 |
| 3 | obando | bulacan | success | 09/01/2020 |
Hopefully someone can shed some light on how to tackle this kind of situation.

Answer 1

Try the following using lag()
with cte as
(
select
*,
lag(city) over (order by id) as ncity,
lag(province) over (order by id) as nprovince
from myTable
)
select
id,
coalesce(city, ncity) as city,
coalesce(province, nprovince) as province,
status,
date
from cte
where status = 'success';
output:
| id | city | province | status | date |
| --- | ------ | ------------ | ------- | ---------- |
| 1 | cainta | rizal | success | 02/01/2020 |
| 2 | pasig | metro manila | success | 06/01/2020 |
| 3 | obando | bulacan | success | 09/01/2020 |

Answer 2

Perhaps a window function can help:
SELECT id, city, province, status, date
FROM (SELECT id,
max(city) OVER w AS city,
max(province) OVER w AS province,
status,
date
FROM atable
WINDOW w AS (PARTITION BY id)) AS q
WHERE status = 'success';

Answer 3

You can use the analytical functions here.
SELECT * FROM
(SELECT T.ID, T.CITY, T.PROVINCE,
MAX(CASE WHEN STATUS = 'success' THEN DATE END)
OVER (PARTITION BY ID ORDER BY DATE) AS DATE,
ROW_NUMBER() OVER (PARTITION BY ID ORDER BY DATE) AS RN,
SUM(CASE WHEN STATUS = 'success' THEN 1 ELSE 0 END)
OVER (PARTITION BY ID) AS CNT
FROM YOUR_TABLE T)
WHERE RN = 1 AND CNT > 0
As you have changed the sample data, You can use the GROUP BY as follows:
SELECT ID, MAX(CITY) AS CITY, MAX(PROVINCE) AS PROVINCE,
MAX(CASE WHEN STATUS = 'success' THEN DATE END) AS DATE
FROM YOUR_TABLE
GROUP BY ID
HAVING SUM(CASE WHEN STATUS = 'success' THEN 1 END) > 0

Answer 4

If you want just one row per id, you can use aggregation:
select id, max(city) as city, max(province) as province,
max(date) filter (where status = 'success') as date
from t
group by id
having count(*) filter (where status = 'success') > 0;
Note that if you can have multiple success dates per id, you can put the on the same row using array_agg():
array_agg(date) filter (where status = 'success') as dates