我正在使用 2 个数据帧，作为示例：

DF1是主表，其中一列包含通常具有变量的方程

DF1 <- read.table(text = 
"Unit     Year  Equation
 1        2020  'x+2*y'
 1        2021  'x+2*y'
 1        2022  'x+2*y'
 2        2020  'x'
 3        2020  'max(y^2, y+2*z)'
 3        2021  'max(y^2, y+2*z)'
 4        2020  '5'
 5        2020  '(x/y)+z'",
header = TRUE, stringsAsFactors = FALSE)

DF2是参考或查找表，它为给定特定年份的变量分配值

DF2 <- read.table(text = 
"Year  Variable  Value
 2020  x         10
 2021  x         15.5
 2022  x         50
 2020  y         1
 2021  y         2
 2022  y         3.5
 2020  z         20
 2021  z         34
 2022  z         11",
header = TRUE, stringsAsFactors = FALSE)

目标是匹配 2 个数据框之间的变量和年份，以便在应用 eval(parse(text=Equation)) 或任何类似内容后可以得出下表：

Unit     Year  Equation
 1       2020  12
 1       2021  19.5
 1       2022  57
 2       2020  10
 3       2020  41
 3       2021  70
 4       2020  5
 5       2020  30

目前我正在使用 for 和 if-else 循环来匹配年份并逐行替换变量。它工作正常，但运行它变得非常慢，因为 DF1 可能包含具有多个变量的数千行。我可以使用其他功能来实现相同的输出吗？

编辑- 在提到的循环中添加以帮助进行比较：

library(dplyr)
library(reshape2)
DF2 = dcast(DF2, Year~Variable, value.var='Value')

  #Adding in this line to avoid replacing "x" in "max":
  DF1$Equation = gsub("max","placeholder",DF1$Equation)

for(i in 1:nrow(DF1)) {
  for (j in 1:nrow(DF2)) {
    if (DF1[i,]$Year==DF2[j,]$Year) {
      #Every variable would be declared here:
      DF1[i,]$Equation = gsub("x",DF2[j,]$x,DF1[i,]$Equation)
      DF1[i,]$Equation = gsub("y",DF2[j,]$y,DF1[i,]$Equation)
      DF1[i,]$Equation = gsub("z",DF2[j,]$z,DF1[i,]$Equation)
    }
  }
}
  #Returning the function:
  DF1$Equation = gsub("placeholder","max",DF1$Equation)

Results_DF1 = DF1 %>% rowwise() %>%
              mutate(Equation = eval(parse(text=Equation)))

回答

left_join(DF1, DF2, 'Year') %>%
  pivot_wider(c(Unit,Year,Equation),Variable, values_from = Value) %>%
  rowwise() %>%
  mutate(a = eval(parse(text = Equation)))

   Unit  Year Equation            x     y     z     a
  <int> <int> <chr>           <dbl> <dbl> <dbl> <dbl>
1     1  2020 x+2*y            10     1      20  12  
2     1  2021 x+2*y            15.5   2      34  19.5
3     1  2022 x+2*y            50     3.5    11  57  
4     2  2020 x                10     1      20  10  
5     3  2020 max(y^2, y+2*z)  10     1      20  41  
6     3  2021 max(y^2, y+2*z)  15.5   2      34  70  
7     4  2020 5                10     1      20   5  
8     5  2020 (x/y)+z          10     1      20  30