req_id_mem = ""
req_workflow_mem = ""
collect_state_main = []
collect_state_temp = []
for req_id, req_datetime, req_workflow in zip(df["TICKET_ID"], df["DATETIMESTANDARD"], df["STATUS"]):
if req_id_mem == "" or req_id_mem != req_id:
req_id_mem = req_id
req_workflow_mem = ""
collect_state_temp = []
if req_workflow_mem == "" and req_workflow == "Open" and req_id_mem == req_id:
req_workflow_mem = req_workflow
collect_state_temp.append(req_id)
collect_state_temp.append(req_workflow)
collect_state_temp.append(req_datetime)
if req_workflow_mem == "Open" and req_workflow == "Closed" and req_id_mem == req_id:
req_workflow_mem = req_workflow
collect_state_temp.append(req_workflow)
collect_state_temp.append(req_datetime)
collect_state_main.append(collect_state_temp)
collect_state_temp = []
DataFrame:
TICKET_ID
DATETIMESTANDARD
STATUS
79355138
3.09.2024 11:54 :18 утра
Открыто
79355138
03.09.2024 21:01:12
Открыто
< td>79355138
6.09.2024 16:52:10
Закрыто
79355138
6.09.2024 16:52:12
ОткрытоОткрыто td>
79355138
10.09.2024 16:01:24
Закрыто
79446344
27.08 /2024 13:32:54
Открыто
79446344
11.09.2024 9:40:17
Закрыто
79446344
11.09.2024 9:40:24
Закрыто
79446344
11.09.2024 9:42:14
< td>Открыть
Результат:
- Он определяет первое состояние Открыто для TICKET_ID и ищет ближайший статус Закрыто
- Для каждого случая будет повторяться поиск пары открытия и закрытия (будут учитываться только первое открытие и первое закрытие)
Добавляю также свое решение, когда я недавно перешел на снежинку:
SELECT
FOD.TICKET_ID,
FOD.FIRSTOPENDATETIME AS OPEN_DATETIME,
MIN(NC.DATETIMESTANDARD) AS CLOSED_DATETIME
FROM
(
SELECT
TICKET_ID,
MIN(DATETIMESTANDARD) AS FIRSTOPENDATETIME,
STATUS
FROM
DB.TABLE
WHERE
(
(STATUS IN ('Open') AND EVENT_TYPE IN ('Ticket Open'))
OR STATUS IN ('Closed')
)
GROUP BY
TICKET_ID, STATUS
) AS FOD
LEFT JOIN DB.TABLE AS NC ON FOD.TICKET_ID = NC.TICKET_ID AND NC.STATUS = 'Closed' AND NC.DATETIMESTANDARD > FOD.FIRSTOPENDATETIME
WHERE
FOD.STATUS = 'Open'
GROUP BY
FOD.TICKET_ID, FOD.FIRSTOPENDATETIME
ORDER BY
FOD.TICKET_ID ASC, FOD.FIRSTOPENDATETIME ASC
Подробнее здесь: https://stackoverflow.com/questions/791 ... one-column