Question

我有两个数据框，
aa的事件发生时间为start和stop-每天很多
prbb有我对这些事件感兴趣的时间start和stop-每天每个事件一个start和一个stop。
我想从aa中提取许多事件，这些事件属于prbb

aa <- data.frame(aaletters = c(rep("a",7), rep("b", 3),rep("c", 1)),
         aastart = as.POSIXct(c("2019-05-02 05:06:35","2019-05-02 12:06:35", "2019-05-03 08:15:52", "2019-05-03 09:15:52", "2019-05-06 05:51:37",
                                  "2019-05-06 07:01:37","2019-05-06 09:51:37","2019-05-02 07:15:32", "2019-05-03 12:14:04", "2019-05-06 12:24:37",
                                  "2019-05-02 03:15:32"
         )),
         aastop = as.POSIXct(c("2019-05-02 05:15:30", "2019-05-02 12:15:30", "2019-05-03 08:44:08","2019-05-03 09:44:08", "2019-05-06 06:51:37",
                               "2019-05-06 07:02:37","2019-05-06 10:02:37","2019-05-02 08:15:32", "2019-05-03 13:41:16", "2019-05-06 13:24:43",
                               "2019-05-02 03:35:32"
                                  )))

prbb <- data.frame(prbbletters = c(rep("a", 3), rep("b", 3), rep("c",3)), 
         prstart = as.POSIXct(c("2019-05-02 06:06:35", "2019-05-03 06:15:52", "2019-05-06 07:51:37", "2019-05-02 06:15:32", "2019-05-03 08:14:04", 
                                "2019-05-06 06:24:37","2019-05-02 06:14:19", "2019-05-03 06:41:35", "2019-05-06 06:17:50"
                                  )),
         prstop = as.POSIXct(c("2019-05-02 23:18:30", "2019-05-03 20:44:08", "2019-05-06 22:37:20", "2019-05-02 23:24:27", "2019-05-03 19:41:16", 
                               "2019-05-06 23:24:43","2019-05-02 19:50:52", "2019-05-03 23:57:47", "2019-05-06 23:56:39"
                                  )))

我尝试了此操作，但没有成功，它忽略了by=

setDT(aa)
setDT(prbb)
aa[inrange(aa$aastart, prbb$prstart, prbb$prstop, incbounds = FALSE) & inrange(aa$aastop, prbb$prstart, prbb$prstop, incbounds = FALSE), by = prbletters]

setDT(aa)
setkey(aa, aastart, aastop)
setDT(prbb)
setkey(prbb, prstart, prstop)
foverlaps(aa, prbb, nomatch = NULL, mult = "first")[ ,  by = prbbletters]

我也尝试过fuzzy_joins，但似乎无法正确整合分组。

#      expected result: 7 rows
#  1: a 2019-05-02 12:06:35 2019-05-02 12:15:30
#  2: a 2019-05-03 08:15:52 2019-05-03 08:44:08
#  3: a 2019-05-03 09:15:52 2019-05-03 09:44:08
#  4: a 2019-05-06 09:51:37 2019-05-06 10:02:37
#  5: b 2019-05-02 07:15:32 2019-05-02 08:15:32
#  6: b 2019-05-03 12:14:04 2019-05-03 13:41:16
#  7: b 2019-05-06 12:24:37 2019-05-06 13:24:43

谢谢！

Answer 1

看来您已经接近。.在使用foverlaps之前只需要添加字母作为键即可。

来自foverlaps-help：
by.x和by.y中的最后两列应分别对应于x和y中的开始和结束间隔列。

因此，请设置要重叠连接的所有键，并确保最后两个键是 start 和 end 。

setDT(aa)
setDT(prbb)
setkey(aa, aaletters, aastart, aastop)     # <-- added aalatters as first key !!
setkey(prbb, prbbletters, prstart, prstop) # <-- added prbbletters as key !!

foverlaps(aa, prbb, mult = "first", nomatch = 0L)

#    aaletters             prstart              prstop             aastart              aastop
# 1:         a 2019-05-02 06:06:35 2019-05-02 23:18:30 2019-05-02 12:06:35 2019-05-02 12:15:30
# 2:         a 2019-05-03 06:15:52 2019-05-03 20:44:08 2019-05-03 08:15:52 2019-05-03 08:44:08
# 3:         a 2019-05-03 06:15:52 2019-05-03 20:44:08 2019-05-03 09:15:52 2019-05-03 09:44:08
# 4:         a 2019-05-06 07:51:37 2019-05-06 22:37:20 2019-05-06 09:51:37 2019-05-06 10:02:37
# 5:         b 2019-05-02 06:15:32 2019-05-02 23:24:27 2019-05-02 07:15:32 2019-05-02 08:15:32
# 6:         b 2019-05-03 08:14:04 2019-05-03 19:41:16 2019-05-03 12:14:04 2019-05-03 13:41:16
# 7:         b 2019-05-06 06:24:37 2019-05-06 23:24:43 2019-05-06 12:24:37 2019-05-06 13:24:43

r中特定组的参考时间子集时间

1 个答案: