我有dataset1有两列ID和Application_SubmittedDate。 Application_SubmittedDate列是日期/时间列。
ID Application_SubmittedDate
6972 2001-05-30 16:57:00
6972 2003-03-08 12:30:00
6972 2006-03-22 17:43:00
6972 2003-08-07 20:20:00
6972 2006-07-28 18:28:00
6972 2001-05-25 17:14:00
6972 2003-09-30 00:48:00
6972 2002-06-04 18:11:00
6972 2006-05-06 17:30:00
6972 2003-02-24 16:02:00
6972 2006-09-16 16:29:00
6972 2003-02-12 22:47:00
6972 2002-08-15 23:30:00
6972 2002-08-31 22:32:00
40841 2002-09-27 05:39:00
40841 2002-01-08 09:05:00
40841 2002-10-07 21:04:00
40841 2002-08-17 18:50:00
59547 2003-08-12 10:45:00
59547 2001-02-20 17:02:00
59547 2002-11-05 23:01:00
60861 2003-10-27 14:40:00
63457 2001-12-05 04:16:00
65048 2002-12-16 10:18:00
65048 2003-12-29 17:52:00
65048 2005-02-20 16:58:00
67037 2004-01-01 18:18:00
67037 2006-06-22 01:04:00
67037 2004-07-31 18:30:00
67037 2004-08-04 14:09:00
67037 2005-04-20 18:06:00
67037 2006-06-15 16:55:00
df1 <- structure(list(ID = c(6972L, 6972L, 6972L, 6972L, 6972L, 6972L,
6972L, 6972L, 6972L, 6972L, 6972L, 6972L, 6972L, 6972L, 40841L,
40841L, 40841L, 40841L, 59547L, 59547L, 59547L, 60861L, 63457L,
65048L, 65048L, 65048L, 67037L, 67037L, 67037L, 67037L, 67037L,
67037L), Application_SubmittedDate = structure(c(991241820, 1047126600,
1143049380, 1060287600, 1154111280, 990810840, 1064882880, 1023214260,
1146936600, 1046102520, 1158424140, 1045090020, 1029454200, 1030833120,
1033105140, 1010480700, 1034024640, 1029610200, 1060685100, 982688520,
1036537260, 1067265600, 1007525760, 1040033880, 1072720320, 1108918680,
1072981080, 1150938240, 1091298600, 1091628540, 1114020360, 1150390500
), class = c("POSIXct", "POSIXt"), tzone = "UTC")), .Names = c("ID",
"Application_SubmittedDate"), class = "data.frame", row.names = c(1L,
18L, 35L, 52L, 69L, 86L, 103L, 137L, 154L, 188L, 205L, 239L,
256L, 273L, 290L, 300L, 305L, 310L, 315L, 327L, 339L, 351L, 352L,
353L, 359L, 371L, 389L, 400L, 411L, 422L, 466L, 477L))
第二个数据集有三列ID，Application_ProcessStartDate，Application_ProcessEndDate。这两个Applicateion ProcessStarDate和EndDate列是日期/时间列。
ID Application_ProcessStartDate Application_ProcessEndDate
65048 2005-02-20 12:44:22 2005-02-23 06:07:45
65048 2006-06-21 17:31:45 2006-06-24 01:42:41
111993 2006-06-21 17:31:45 2006-06-24 01:42:41
df2 <- structure(list(ID = c(65048L, 65048L, 111993L), Application_ProcessStartDate = structure(c(1108903462,
1150911105, 1150911105), class = c("POSIXct", "POSIXt"), tzone = "UTC"),
Application_ProcessEndDate = structure(c(1109138865, 1151113361,
1151113361), class = c("POSIXct", "POSIXt"), tzone = "UTC")), .Names = c("ID",
"Application_ProcessStartDate", "Application_ProcessEndDate"), row.names = c(NA,
-3L), class = c("tbl_df", "tbl", "data.frame"))
我的目标是合并第一）通过ID 2）在那些ID合并那些来自df1的那些，其中Application_SubmittedDate值介于Application_ProcessStartDate，Application_ProcessEndDate值之间。
最终结果看起来像这样
ID Application_SubmittedDate Application_ProcessStartDate Application_ProcessEndDate
6972 2001-05-30 16:57:00
6972 2003-03-08 12:30:00
6972 2006-03-22 17:43:00
6972 2003-08-07 20:20:00
6972 2006-07-28 18:28:00
6972 2001-05-25 17:14:00
6972 2003-09-30 00:48:00
6972 2002-06-04 18:11:00
6972 2006-05-06 17:30:00
6972 2003-02-24 16:02:00
6972 2006-09-16 16:29:00
6972 2003-02-12 22:47:00
6972 2002-08-15 23:30:00
6972 2002-08-31 22:32:00
40841 2002-09-27 05:39:00
40841 2002-01-08 09:05:00
40841 2002-10-07 21:04:00
40841 2002-08-17 18:50:00
59547 2003-08-12 10:45:00
59547 2001-02-20 17:02:00
59547 2002-11-05 23:01:00
60861 2003-10-27 14:40:00
63457 2001-12-05 04:16:00
65048 2002-12-16 10:18:00
65048 2003-12-29 17:52:00
65048 2005-02-20 16:58:00 2005-02-20 12:44:22 2005-02-23 06:07:45
65048 NA 2006-06-21 17:31:45 2006-06-24 01:42:41
67037 2004-01-01 18:18:00
67037 2006-06-22 01:04:00
67037 2004-07-31 18:30:00
67037 2004-08-04 14:09:00
67037 2005-04-20 18:06:00
67037 2006-06-15 16:55:00
111993 NA 2006-06-21 17:31:45 2006-06-24 01:42:41
我试过foverlaps这不处理日期/时间值，只有日期值，所以这是排除。我也试过了sqldf图书馆的JOIN，但这只是INNER JOINS，而不是OUTER JOINS所以这也被排除了。不知道如何做到这一点。非常感谢任何帮助或建议。
使用data.table可以实现另一种解决方案。该方法将加入df1和df2为right_join和left_join，然后合并两者。
library(data.table)
setDT(df1)
setDT(df2)
rhs_join <- df1[df2,
.(i.ID, x.Application_SubmittedDate, i.Application_ProcessStartDate,
i.Application_ProcessEndDate),
on = .(ID = ID, Application_SubmittedDate >= Application_ProcessStartDate,
Application_SubmittedDate <= Application_ProcessEndDate)][,.(ID = i.ID, Application_SubmittedDate = x.Application_SubmittedDate,
Application_ProcessStartDate = i.Application_ProcessStartDate,
Application_ProcessEndDate = i.Application_ProcessEndDate)]
lhs_join <- df2[df1,
.(i.ID, Application_SubmittedDate, x.Application_ProcessStartDate,
x.Application_ProcessEndDate),
on = .(ID = ID, Application_ProcessStartDate <= Application_SubmittedDate,
Application_ProcessEndDate >= Application_SubmittedDate)][,.(ID = i.ID, Application_SubmittedDate = Application_SubmittedDate,
Application_ProcessStartDate = x.Application_ProcessStartDate,
Application_ProcessEndDate = x.Application_ProcessEndDate)]
#Merge both data.frames
merge(rhs_join, lhs_join, all=TRUE)
ID Application_SubmittedDate Application_ProcessStartDate Application_ProcessEndDate
1: 6972 2001-05-25 17:14:00 <NA> <NA>
2: 6972 2001-05-30 16:57:00 <NA> <NA>
3: 6972 2002-06-04 18:11:00 <NA> <NA>
.....
.....
.....
23: 63457 2001-12-05 04:16:00 <NA> <NA>
24: 65048 <NA> 2006-06-21 17:31:45 2006-06-24 01:42:41
25: 65048 2002-12-16 10:18:00 <NA> <NA>
26: 65048 2003-12-29 17:52:00 <NA> <NA>
27: 65048 2005-02-20 16:58:00 2005-02-20 12:44:22 2005-02-23 06:07:45
28: 67037 2004-01-01 18:18:00 <NA> <NA>
29: 67037 2004-07-31 18:30:00 <NA> <NA>
30: 67037 2004-08-04 14:09:00 <NA> <NA>
31: 67037 2005-04-20 18:06:00 <NA> <NA>
32: 67037 2006-06-15 16:55:00 <NA> <NA>
33: 67037 2006-06-22 01:04:00 <NA> <NA>
34: 111993 <NA> 2006-06-21 17:31:45 2006-06-24 01:42:41
问题中的描述似乎不清楚，但也许你想要其中一个左连接。对于问题中显示的数据，这些数据分别产生32行和3行。
library(sqldf)
sqldf("select a.*,
b.Application_ProcessStartDate,
b.Application_ProcessEndDate
from df1 a left join df2 b
on a.ID = b.ID and
a.Application_SubmittedDate between
b.Application_ProcessStartDate and
b.Application_ProcessEndDate")
sqldf("select a.*,
b.Application_ProcessStartDate,
b.Application_ProcessEndDate
from df2 b left join df1 a
on a.ID = b.ID and
a.Application_SubmittedDate between
b.Application_ProcessStartDate and
b.Application_ProcessEndDate")
或者你可能正在寻找两者的结合：
sqldf("select a.*,
b.Application_ProcessStartDate,
b.Application_ProcessEndDate
from df1 a left join df2 b
on a.ID = b.ID and
a.Application_SubmittedDate between
b.Application_ProcessStartDate and
b.Application_ProcessEndDate
union
select a.*,
b.Application_ProcessStartDate,
b.Application_ProcessEndDate
from df2 b left join df1 a
on a.ID = b.ID and
a.Application_SubmittedDate between
b.Application_ProcessStartDate and
b.Application_ProcessEndDate")