R：合并2个数据帧，并在前一行和后一行进行索引

hachondeoro 发表于 Dev

Hachondeoro

可以说我有一个像这样的数据集：

origin=data.frame(Date=as.Date(c("2016-08-05","2016-08-04","2016-08-03")),
                  L=c(1,2,3),
                  Type=c("H","L","H"))

        Date L Type
1 2016-08-05 1    H
2 2016-08-04 2    L
3 2016-08-03 3    H

end=data.frame(Date=as.Date(c("2016-08-05","2016-08-04","2016-08-03","2016-08-02","2016-08-01")),
               N=c(50,40,30,20,10),
               Name=c("CA","CB","CC","CD","CE"),
               Vol=c(2,1,2,2,3),
               Act=c(0.1,0.2,0.3,0.2,0.2))

        Date  N Name Vol Act
1 2016-08-05 50   CA   2 0.1
2 2016-08-04 40   CB   1 0.2
3 2016-08-03 30   CC   2 0.3
4 2016-08-02 20   CD   2 0.2
5 2016-08-01 10   CE   3 0.2

我想要这样的东西：

        Date L Type  N Name Vol Act
3 2016-08-05 1    H 50   CA   2 0.1
3 2016-08-05 1    H 40   CB   1 0.2
3 2016-08-05 1    H 30   CC   2 0.3
2 2016-08-04 2    L 40   CB   1 0.2
2 2016-08-04 2    L 30   CC   2 0.3
2 2016-08-04 2    L 20   CD   2 0.2
1 2016-08-03 3    H 30   CC   2 0.3
1 2016-08-03 3    H 20   CD   2 0.2
1 2016-08-03 3    H 10   CE   3 0.2

我想保留“起源”的原始列日期，在合并中，我想将其与“结束”的当前日期和以前的日期值（两个先前的值）合并，就像用循环合并一样。在其他帖子中，仅匹配公共值，这将给出3行结果：

merge(x = origin, y = end, by = "Date")

        Date L Type  N Name Vol Act
1 2016-08-03 3    H 30   CC   2 0.3
2 2016-08-04 2    L 40   CB   1 0.2
3 2016-08-05 1    H 50   CA   2 0.1

这是非常不同的，并且不会按当前行值和上一行值合并两个数据帧，因此我无法弄清楚该如何进行。

智慧

看起来像是foverlaps从data.table对这份工作西装：

# prepare data and add extra columns for foverlaps join which relies on columns instead of one
library(data.table)
setDT(origin)[, DateStart := Date - 2]
setDT(end)[, DateStart := Date]
setkey(origin, DateStart, Date)

# join two tables with foverlaps and remove subsidiary columns
foverlaps(end, origin, type = "within")[, `:=` (DateStart = NULL, i.Date = NULL, i.DateStart = NULL)][order(Date)]

#          Date L Type  N Name Vol Act
# 1: 2016-08-03 3    H 30   CC   2 0.3
# 2: 2016-08-03 3    H 20   CD   2 0.2
# 3: 2016-08-03 3    H 10   CE   3 0.2
# 4: 2016-08-04 2    L 40   CB   1 0.2
# 5: 2016-08-04 2    L 30   CC   2 0.3
# 6: 2016-08-04 2    L 20   CD   2 0.2
# 7: 2016-08-05 1    H 50   CA   2 0.1
# 8: 2016-08-05 1    H 40   CB   1 0.2
# 9: 2016-08-05 1    H 30   CC   2 0.3

或使用version的non-equi加入功能：data.table1.9.7

setDT(origin)[, `:=` (DateEnd = Date, StartDate = Date - 2)]
             [setDT(end), on = .(DateEnd >= Date, StartDate <= Date), allow = T]

#          Date L Type    DateEnd  StartDate  N Name Vol Act
# 1: 2016-08-05 1    H 2016-08-05 2016-08-05 50   CA   2 0.1
# 2: 2016-08-04 2    L 2016-08-04 2016-08-04 40   CB   1 0.2
# 3: 2016-08-05 1    H 2016-08-04 2016-08-04 40   CB   1 0.2
# 4: 2016-08-03 3    H 2016-08-03 2016-08-03 30   CC   2 0.3
# 5: 2016-08-04 2    L 2016-08-03 2016-08-03 30   CC   2 0.3
# 6: 2016-08-05 1    H 2016-08-03 2016-08-03 30   CC   2 0.3
# 7: 2016-08-03 3    H 2016-08-02 2016-08-02 20   CD   2 0.2
# 8: 2016-08-04 2    L 2016-08-02 2016-08-02 20   CD   2 0.2
# 9: 2016-08-03 3    H 2016-08-01 2016-08-01 10   CE   3 0.2

删除辅助列应该很简单。

本文收集自互联网，转载请注明来源。

如有侵权，请联系[email protected] 删除。