sql表有出租车司机的数据，列是：

id  integer
client_id   integer (Foreign keyed to events.rider_id)
driver_id   integer
city_id integer (Foreign keyed to cities.city_id)
client_rating   integer
driver_rating   integer
predicted_eta   integer
actual_eta  integer
first_completed_date Timestamp ,
status  Enum(‘completed’, ‘cancelled_by_driver’, ‘cancelled_by_client’)

字符串
如何计算过去30天内所有已完成的行程的实际预计到达时间与预计到达时间之间的第90百分位数差异？

select

  percentile_cont(0.90) within group (order by actual_eta-predicted_eta) as percentile_90_diff

from trips t where status='completed'

and first_completed_date > (first_completed_date - INTERVAL '30 DAY')::DATE

型
我关心的是最后一部分- first_completed_date > -在过去的30天内。从first_completed_date开始计算过去的30天是否正确？