使用特定值顺序对数据表进行排序

Question

这是一个简短的data.table:

DT <- data.table(Tag1 = c(22,253,6219,6219,252862,252864,312786,312812),
                 Tag2 = c(22,255,6220,252857,252863,252865,251191,252863),
                 Date= as.Date(as.character(c("7/25/2008","6/15/2000","6/30/2000","9/6/2002","9/6/2002","9/6/2002","9/3/2003","9/5/2003")),format = "%m/%d/%Y"))
DT

     Tag1   Tag2       Date
1:     22     22 2008-07-25
2:    253    255 2000-06-15
3:   6219   6220 2000-06-30
4:   6219 252857 2002-09-06
5: 252862 252863 2002-09-06
6: 252864 252865 2002-09-06
7: 312786 251191 2003-09-03
8: 312812 252863 2003-09-05

我想按3列升序对data.table进行排序:Tag1,Tag2和Date.我测试过:

> test <- DT[order(Tag1, Tag2, Date)]
> test
     Tag1   Tag2       Date
1:     22     22 2008-07-25
2:    253    255 2000-06-15
3:   6219   6220 2000-06-30
4:   6219 252857 2002-09-06
5: 252862 252863 2002-09-06
6: 252864 252865 2002-09-06
7: 312786 251191 2003-09-03
8: 312812 252863 2003-09-05

但是,我想对data.table进行排序,如下所示:

> test
         Tag1   Tag2       Date
    1:     22     22 2008-07-25
    2:    253    255 2000-06-15
    3:   6219   6220 2000-06-30
    4:   6219 252857 2002-09-06
    5: 252862 252863 2002-09-06
    6: 312812 252863 2003-09-05
    7: 252864 252865 2002-09-06
    8: 312786 251191 2003-09-03

特别是,Tag1或Tag1的重复值应该一个接一个地放置(例如:Tag1为6219,Tag2为252863).我怎样才能做到这一点 ？

编辑:

建议的解决方案适用于简短的data.table(如上面的data.table).这是一个更长的版本:

DT <- data.table(Tag1 = c(252860, 252862, 312812, 252864, 252866, 252868, 252870, 318880, 252872, 252874, 252876, 252878, 252880, 252880, 252881, 252883,
252885, 252887, 311264, 252889, 252889, 252892, 318879, 318880, 318881), Tag2 = c(252861, 252863, 252863, 252865, 252867, 252869, 252871, 252871, 252873,
252875, 252877, 252879, 414611, 905593, 252882, 252884, 252886, 252888, 252888, 252890, 318904, 252893, 318878, 414547, 318882), Date = c("9/6/2002",
"9/6/2002", "9/5/2003", "9/6/2002", "9/6/2002", "9/6/2002", "9/6/2002", "10/8/2003", "9/6/2002", "9/6/2002", "9/6/2002", "9/6/2002", "10/5/2004",
"9/6/2002", "9/6/2002", "9/6/2002", "9/10/2002", "9/10/2002", "7/15/2003", "9/10/2002", "10/15/2003", "9/10/2002", "10/8/2003", "9/29/2004","10/8/2003"))

这是预期的结果(即data.table"After").特别是,data.table"After"应该遵循两个条件:

1)行按日期按升序排序

2)Tag1或Tag1的重复值一个接一个地放置(最终不需要按升序排列)

Tag1和Tag2的所有重复值均为黄色.

Answer 1

旧秩序

df[order(Tag1, Tag2, Date)]
#      Tag1   Tag2       Date
# 1:     22     22 2008-07-25
# 2:    253    255 2000-06-15
# 3:   6219   6220 2000-06-30
# 4:   6219 252857 2002-09-06
# 5: 252862 252863 2002-09-06
# 6: 252864 252865 2002-09-06
# 7: 312786 251191 2003-09-03
# 8: 312812 252863 2003-09-05

新订单
排序Date列按降序排列,然后Tag1按升序排序Tag2.

setcolorder(dt1 <- df[order(-Date)][order(Tag1), .SD, by = Tag2], colnames(df))

dt1
#      Tag1   Tag2       Date
# 1:     22     22 2008-07-25
# 2:    253    255 2000-06-15
# 3:   6219 252857 2002-09-06
# 4:   6219   6220 2000-06-30
# 5: 252862 252863 2002-09-06
# 6: 312812 252863 2003-09-05
# 7: 252864 252865 2002-09-06
# 8: 312786 251191 2003-09-03

评论中@akrun的解决方案扰乱了数据的结构.这是比较.看看#4:6219应该有252857而不是251191

df[,lapply(df, sort)]
#      Tag1   Tag2       Date
# 1:     22     22 2000-06-15
# 2:    253    255 2000-06-30
# 3:   6219   6220 2002-09-06
# 4:   6219 251191 2002-09-06
# 5: 252862 252857 2002-09-06
# 6: 252864 252863 2003-09-03
# 7: 312786 252863 2003-09-05
# 8: 312812 252865 2008-07-25