R语言 使用输入向量创建多个新列

yuvru6vn  于 2023-11-14  发布在  其他
关注(0)|答案(1)|浏览(105)

我有大量的数据集，我需要计算滞后的数字。结果数据集需要为每一天，列的数量'回头看'的要求。在我的例子中，我看了5天切片为10个以前的时期。我试图找出如何推广这一点，以涵盖任何数量的时间片长度和数量的要求回头看（例如，前5年为365天）。
x1c 0d1x的数据
我已经包括了一个60天的单站和3个代码示例。第一个示例显示了通过硬编码来实现。如果我有一百个回顾期，这会很乏味。
第二个例子展示了我如何开始使用参数，这是值得的。
第三（这不起作用）显示了我希望如何编写某种语句，该语句在要创建的度量向量之间发生变化，使用第二个向量作为滞后量。也许使用{col}而不是在向量中预先指定名称更好？我只是不知道如何引用一个“固定”变量（第一个汇总，但创建一个动态数量的新变量连续较长的时期。
谢谢！

  1. pacman::p_load(tidyverse, RcppRoll)
  3. # The input dataset
  4. Daily <- structure(list(stnID = c(165638621686, 165638621686, 165638621686, 
  5. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  6. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  7. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  8. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  9. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  10. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  11. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  12. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  13. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  14. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  15. 165638621686, 165638621686, 165638621686, 165638621686, 165638621686, 
  16. 165638621686, 165638621686), date = structure(c(10957, 10958, 
  17. 10959, 10960, 10961, 10962, 10963, 10964, 10965, 10966, 10967, 
  18. 10968, 10969, 10970, 10971, 10972, 10973, 10974, 10975, 10976, 
  19. 10977, 10978, 10979, 10980, 10981, 10982, 10983, 10984, 10985, 
  20. 10986, 10987, 10988, 10989, 10990, 10991, 10992, 10993, 10994, 
  21. 10995, 10996, 10997, 10998, 10999, 11000, 11001, 11002, 11003, 
  22. 11004, 11005, 11006, 11007, 11008, 11009, 11010, 11011, 11012, 
  23. 11013, 11014, 11015, 11016), class = "Date"), measure = c(15.3333333333333, 
  24. 15.9791666666667, 16.65, 12.675, 9.32916666666667, 9.71041666666667, 
  25. 11.8916666666667, 11.9958333333333, 11.025, 10.94375, 11.3791666666667, 
  26. 9.04166666666667, 10.5604166666667, 10.8583333333333, 11.4083333333333, 
  27. 10.1979166666667, 10.19375, 13.1645833333333, 13.7604166666667, 
  28. 13.21875, 11.16875, 10.43125, 11.0604166666667, 13.4041666666667, 
  29. 14.0979166666667, 10.8521739130435, 8.54375, 5.44375, 8.06666666666667, 
  30. 9.77291666666667, 10.2676470588235, 11.5979166666667, 12.1375, 
  31. 11.7958333333333, 12.3916666666667, 12.7875, 13.4604166666667, 
  32. 10.7541666666667, 10.1979166666667, 10.9145833333333, 11.76875, 
  33. 13.6291666666667, 12.5, 10.9416666666667, 12.16875, 12.2229166666667, 
  34. 12.0541666666667, 11.69375, 11.05, 12.3229166666667, 12.1208333333333, 
  35. 11.5020833333333, 13.1770833333333, 11.3833333333333, 9.88, 10.9520833333333, 
  36. 11.275, 11.4208333333333, 11.3270833333333, 11.0104166666667)), class = c("tbl_df", 
  37. "tbl", "data.frame"), row.names = c(NA, -60L))
  39. # The hand-code approach
  41.     rolled1 <- Daily %>% group_by(stnID) %>% 
  42.         mutate(measure_01 := roll_meanr(measure, 5, na.rm=TRUE)) %>% 
  43.     
  44.         mutate(measure_02 := lag(measure_01, 10)) %>% 
  45.         mutate(measure_03 := lag(measure_01, 15)) %>% 
  46.         mutate(measure_04 := lag(measure_01, 20)) %>% 
  47.         mutate(measure_05 := lag(measure_01, 25)) %>% 
  48.         mutate(measure_06 := lag(measure_01, 30)) %>% 
  49.         mutate(measure_07 := lag(measure_01, 35)) %>% 
  50.         mutate(measure_08 := lag(measure_01, 40)) %>% 
  51.         mutate(measure_09 := lag(measure_01, 45)) %>% 
  52.         mutate(measure_10 := lag(measure_01, 50)) 
  53.     
  54. # The parameter approach
  55.     
  56.     
  57. periodN    <- 10 # number of lagged periods we want to create
  58. periodSize <- 5  # of days back to look
  60. varList <- paste0("measure_",sprintf("%02d", seq_along(1:periodN)))
  61.     
  62.     rolled2 <- Daily %>% group_by(stnID) %>% 
  63.         mutate(!!varList[1] := roll_meanr(measure, periodSize, na.rm=TRUE)) %>% 
  64.     
  65.         mutate(!!varList[2] := lag(measure_01, periodSize * 2)) %>% 
  66.         mutate(!!varList[3] := lag(measure_01, periodSize * 3)) %>% 
  67.         mutate(!!varList[4] := lag(measure_01, periodSize * 4)) %>% 
  68.         mutate(!!varList[5] := lag(measure_01, periodSize * 5)) %>% 
  69.         mutate(!!varList[6] := lag(measure_01, periodSize * 6)) %>% 
  70.         mutate(!!varList[7] := lag(measure_01, periodSize * 7)) %>% 
  71.         mutate(!!varList[8] := lag(measure_01, periodSize * 8)) %>% 
  72.         mutate(!!varList[9] := lag(measure_01, periodSize * 9)) %>% 
  73.         mutate(!!varList[10]:= lag(measure_01, periodSize * 10))
  74.     
  75. # the vector approach (I have no idea what I am doing here)
  77. varList <- paste0("measure_",sprintf("%02d", seq(2, periodN, 1)))
  78. numList <- periodSize * seq(2, periodN, 1) 
  80.     # I want to loop over the vectors which start from measure_02 to measure_10, and 10 to 50
  82.     rolled3 <- Daily %>% group_by_stnID) %>% 
  83.                 mutate(!!varList[1] := roll_meanr(measure, periodSize, na.rm=TRUE)) %>% 
  84.         mutate(across(varList) ~ lag(measure01, numList))

字符串
我用前两种方法都很成功，我查过例子，但似乎找不到一个能做到这一点的。
我是否应该先“创建”结果列，在每个列中放入滞后量，然后进行突变，使每个列成为滞后[目标日期，.x]？
我头好痛。

r

1条答案

按热度按时间
41zrol4v

41zrol4v1#

从这个答案中提取的。这是你想要的吗？

  1. library(data.table)
  2. library(rlang)
  3. library(tidyverse)
  5. n_lags <- 50
  6. step <- 5
  8. lags <- function(var, n = 50, step = 5) {
  9.   var <- enquo(var)
  10.   indices <- seq(5, n, by = 5)
  11.   
  12.   # create a list of quosures by looping over `indices`
  13.   # then give them names for `mutate` to use later
  14.   map(indices, ~ quo(lag(!!var, !!.x))) %>%
  15.     set_names(sprintf("measure_%02d", indices))
  16. }
  18. # unquote the list of quosures so that they are evaluated by `mutate`
  19. Daily %>% 
  20.   group_by(stnID) %>%
  21.   mutate(measure_01 = frollmean(measure, n = 5, na.rm=TRUE)) %>% 
  22.   mutate_at(vars(measure_01), funs(!!!lags(measure_01, n_lags, step)))
  23. #> 
  24. #> # A tibble: 60 × 14
  25. #>           stnID date       measure measure_01 measure_05 measure_10 measure_15
  26. #>           <dbl> <date>       <dbl>      <dbl>      <dbl>      <dbl>      <dbl>
  27. #>  1 165638621686 2000-01-01   15.3        NA         NA           NA         NA
  28. #>  2 165638621686 2000-01-02   16.0        NA         NA           NA         NA
  29. #>  3 165638621686 2000-01-03   16.6        NA         NA           NA         NA
  30. #>  4 165638621686 2000-01-04   12.7        NA         NA           NA         NA
  31. #>  5 165638621686 2000-01-05    9.33       14.0       NA           NA         NA
  32. #>  6 165638621686 2000-01-06    9.71       12.9       NA           NA         NA
  33. #>  7 165638621686 2000-01-07   11.9        12.1       NA           NA         NA
  34. #>  8 165638621686 2000-01-08   12.0        11.1       NA           NA         NA
  35. #>  9 165638621686 2000-01-09   11.0        10.8       NA           NA         NA
  36. #> 10 165638621686 2000-01-10   10.9        11.1       14.0         NA         NA
  37. #> # ℹ 50 more rows
  38. #> # ℹ 7 more variables: measure_20 <dbl>, measure_25 <dbl>, measure_30 <dbl>,
  39. #> #   measure_35 <dbl>, measure_40 <dbl>, measure_45 <dbl>, measure_50 <dbl>

字符串
创建于2023-10-28使用reprex v2.0.2

展开查看全部
赞(0）分享  回复(0）举报  2023-11-14
我来回答

相关问题

    查看更多

    热门标签

    更多
    JavaquerypythonNode开发语言requestUtil数据库Table后端算法LoggerMessageElementParser

    最新问答

    更多
    • 蜀ICP备13028337号-1 大数据知识库 https://www.saoniuhuo.com © All rights reserved
    • 本站内容来源互联网,如果侵犯您的权益请联系我们删除, 联系方式：448109455@qq.com