使用lubridate的floor_date函数:

df$interv<-lubridate::floor_date(df$time, unit="15 minutes")

(这假设您的时间列是某种公认的日期时间格式(例如POSIXt)，似乎与您的样本格式无关.)

(This assumes that your time column is some recognized date-time format (e.g., POSIXt), which it doesn't seem to be from your sample's formatting.)

例如:

a<-Sys.time()
a
[1] "2019-01-08 09:22:22 EST"
floor_date(a,unit="15 minutes")
[1] "2019-01-08 09:15:00 EST"