I have table like this
user_id | date | time|
"1" | "2017-01-03"| "06:59:35"
"1" | "2017-01-03"| "07:01:17"
"1" | "2017-01-03"| "12:03:21"
"1" | "2017-01-03"| "16:06:14"
"2" | "2017-01-03"| "07:10:52"
"2" | "2017-01-03"| "07:11:38"
"2" | "2017-01-03"| "07:12:04"
"3" | "2017-01-03"| "07:12:06"
"3" | "2017-01-03"| "09:12:33"
"3" | "2017-01-03"| "16:13:29"
This is my mysql query
SELECT
col_user_id as user_id,
col_date as tanggal,
if(MAKETIME(HOUR(MIN(col_jam),MINUTE(MIN(col_jam),00) <= '12:00', MAKETIME(HOUR(MIN(col_jam),MINUTE(MIN(col_jam),00), NULL) as jam_masuk,
if(MAKETIME(HOUR(MAX(col_jam),MINUTE(MAX(col_jam),00) > '12:00', MAKETIME(HOUR(MAX(col_jam),MINUTE(MAX(col_jam),00), NULL) as jam_keluar
FROM tb_kehadiran_temp
WHERE
col_date >= LAST_DAY(CURRENT_DATE) + INTERVAL 1 DAY - INTERVAL 2 MONTH
AND
col_date < LAST_DAY(CURRENT_DATE) + INTERVAL 1 DAY
GROUP BY col_user_id,col_date
I want my result like this on postgre query:
user_id | date | in | out
1 | 2017-01-03 | 06:59:35 | 16:06:14
2 | 2017-01-03 | 07:10:52 | null
3 | 2017-01-03 | 07:12:04 | 16:13:29
and i want to move to other table with condition insert or update if data is exists
I think a simple GROUP BY
query should work here, assuming your time column is an actual time type, or if it's text, it is fixed width:
SELECT
user_id,
date,
MIN(time) AS "in", -- DON'T use in to name a column; it's a keyword
CASE WHEN MAX(time) < '12:00:00' THEN NULL ELSE MAX(time) END AS out
FROM tb_kehadiran_temp
GROUP BY
user_id,
date;
Edit:
If you needed to populate another table based on the results of the above query, you could try using INSERT INTO ... SELECT
syntax, e.g.
INSERT INTO otherTable (user_id, date, min_time, max_time)
SELECT
user_id,
date,
MIN(time) AS "in",
CASE WHEN MAX(time) < '12:00:00' THEN NULL ELSE MAX(time) END AS out
FROM tb_kehadiran_temp
GROUP BY
user_id,
date;