R：根据现有数据框上的多个条件添加数据的列和行

Question

我想重组我的土地利用分类数据框，并根据数据框的条件添加新的行和列。我一直在使用 dplyr 进行尝试，但是我发现的示例倾向于减少列或行，而不是根据条件增加行数。我试图遍历数据集以添加行，但想知道在 dplry 中是否有更好的方法？我也愿意使用不同的库，但它的分类数据集非常大，dplyr 似乎可以很好地与数据框配合使用？

这是我当前数据框的代码示例 (df_old) 以及我希望它最终的样子 (df_new)。

我想要做的是，每次 Year1990-2015 更改它都会创建一个新行。示例：ID 424，在 1990 年为 51，但在 2000 年更改为 21 并保持 21 直到今天。这意味着 ID 424 的新数据框应该有两行。一个标有 Start_Year 的标签表示 1990 年土地利用的开始为 Forest (Landuse = 51)，并且在 2000 年发生变化之前一直是 Forest。由于 2000 年它是 Pavement，我们假设它在 1999 年仍然是 Forest，而 End_Year 将是 1999 年ID 424 的第一行。然后，ID 424 出现一个新行，其中 Start_Year 为 2000，因为它更改为 Pavement (Landuse = 21)，并且在 End_year（现在）之前保持 21。

为了添加上下文，数据集表示城市中区域的变化情况，其中 1990-2015 年的数字用于识别不同的土地利用分类（21 = 路面，24 = 公园，25 = 住宅，51 = 森林，41 = 农业）。

df_old <- data.frame(ID = c(424,426,427,428),
             Parameter= c(0.01,0.03,0.03,0.01),
             City = c("Abbotsford","Abbotsford","Abbotsford","Abbotsford"),
             Area = c(3.12,7.98,2.01,0.48),
             Year1990 = c(51,51,51,41),
             Year2000 = c(21,51,51,41),
             Year2005 = c(21,51,51,25),
             Year2010 = c(21,51,51,24),
             Year2015 = c(21,51,51,25))

df_new <- data.frame(ID = c(424,424,426,427,428,428,428,428),
             Parameter= c(0.01,0.01,0.03,0.03,0.01,0.01,0.01,0.01),
             City = c("Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford"),
             Area = c(3.12,3.12,7.98,2.01,0.48,0.48,0.48,0.48),
             Start_Year = c(1990,2000,1990,1990,1990,2005,2010,2015),
             End_Year = c(1999,"present","present","present",2004,2009,2014,"present"),
             Landuse = c("51-51","51-21","51-51","51-51","41-41","41-25","25-24","24-25"))

这就是我想要的最终产品：

Answer 1

此解决方案适用于您的示例数据，但很难确定管理您所需操作的“规则”（因此很难知道它是否适用于您的真实数据）。如果您的真实数据失败，请使用更多信息编辑您的帖子。

library(tidyverse)

df_old <- data.frame(ID = c(424,426,427,428),
                     Parameter= c(0.01,0.03,0.03,0.01),
                     City = c("Abbotsford","Abbotsford","Abbotsford","Abbotsford"),
                     Area = c(3.12,7.98,2.01,0.48),
                     Year1990 = c(51,51,51,41),
                     Year2000 = c(21,51,51,41),
                     Year2005 = c(21,51,51,25),
                     Year2010 = c(21,51,51,24),
                     Year2015 = c(21,51,51,25))

df_new <- data.frame(ID = c(424,424,426,427,428,428,428,428),
                     Parameter= c(0.01,0.01,0.03,0.03,0.01,0.01,0.01,0.01),
                     City = c("Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford","Abbotsford"),
                     Area = c(3.12,3.12,7.98,2.01,0.48,0.48,0.48,0.48),
                     Start = c(1990,2000,1990,1990,1990,2005,2010,2015),
                     End = c(1999,"present","present","present",2004,2009,2014,"present"),
                     LU = c("51-51","51-21","51-51","51-51","41-41","41-25","25-24","24-25"))


df_old %>%
  pivot_longer(cols = -c(1:4)) %>%
  group_by(ID) %>%
  mutate(Start = as.numeric(str_extract(name, "\\d+"))) %>%
  mutate(`LU-LU` = paste(lag(value, default = max(value)), "-", value, sep = "")) %>%
  distinct(`LU-LU`, .keep_all = TRUE) %>%
  group_by(ID) %>%
  filter(value != lag(value, default = 0)) %>%
  group_by(ID) %>%
  mutate(End = lead(Start, default = NA) - 1,
         End = replace_na(End, "present")) %>%
  select(c(ID, Parameter, City, Area, Start, End, `LU-LU`))
#> # A tibble: 8 × 7
#> # Groups:   ID [4]
#>      ID Parameter City        Area Start End     `LU-LU`
#>   <dbl>     <dbl> <chr>      <dbl> <dbl> <chr>   <chr>  
#> 1   424      0.01 Abbotsford  3.12  1990 1999    51-51  
#> 2   424      0.01 Abbotsford  3.12  2000 present 51-21  
#> 3   426      0.03 Abbotsford  7.98  1990 present 51-51  
#> 4   427      0.03 Abbotsford  2.01  1990 present 51-51  
#> 5   428      0.01 Abbotsford  0.48  1990 2004    41-41  
#> 6   428      0.01 Abbotsford  0.48  2005 2009    41-25  
#> 7   428      0.01 Abbotsford  0.48  2010 2014    25-24  
#> 8   428      0.01 Abbotsford  0.48  2015 present 24-25

^{由reprex package (v2.0.1) 于 2021-12-03 创建}