如何在 R 中创建一个新列来匹配来自两个不同数据框中的多个值

Question

I have 2 data frames with thousands of variables.

One has students of different ages and the different teachers that evaluated them. All teachers evaluated multiple different students but not every student.

Teacher Student Age
0123    1       7
0145    1       7
0163    1       7
0175    2       8
0123    2       8
0194    2       8
0123    3       7 
0145    3       7

Then I have the teacher's ratings for specific stereotypes regarding the different ages. Each teacher made on rating for each age group stereotype. The data frame looks like this.

Teacher Age 7   Age 8  Age 9
0123    1       1      1
0145    5       7      3
0163    4       7      1
0175    6       8      1
0183    3       8      1
0194    2       8      1
0120    3       7      4

I want to create a new column in the first data frame where the teachers in each row are matched, and the values are their stereotype response depending on the age of each student. For example, in this new column, the value in the first row would be teacher 123's stereotype response for 7 year olds. In this case that is a 1.

Thank you so much for your help. I'm new to R and I have no idea where to start with this.

Edit: I would like the output to look like this:

Teacher Student Age AgeStereotype
0123    1       7   1
0145    1       7   5
0163    1       7   4
0175    2       8   8
0123    2       8   1
0194    2       8   8
0123    3       7   1
0145    3       7   5

Answer 1

AS <- apply(DF1[,c("Teacher", "Age")], 1, function(x) {
    DF2[which(DF2$Teacher == x[1]), which(grepl(x[2], names(DF2)))]
    })
DF1["AgeStereotype"] <- AS

with DF1 and DF2 your first and second data frames, respectively.

Output:

  Teacher Student Age AgeStereotype
1     123       1   7             1
2     145       1   7             5
3     163       1   7             4
4     175       2   8             8
5     123       2   8             1
6     194       2   8             8
7     123       3   7             1
8     145       3   7             5

Answer 2

You could use [ ie:

transform(df1,AgeStereotype = `rownames<-`(df2,df2$Teacher)[cbind(Teacher,paste("Age",Age))])

  Teacher Student Age AgeStereotype
1     123       1   7             1
2     145       1   7             5
3     163       1   7             4
4     175       2   8             8
5     123       2   8             1
6     194       2   8             8
7     123       3   7             1
8     145       3   7             5

Answer 3

This task is best solved by transforming your second dataframe to a long dataframe and then joining it to your first dataframe. Many ways exist to accomplish this in R , here is a clean way to do it within the tidyverse , specifically with dplyr and tidyr functions.

# Recreating your data
df1 <- tibble::tribble(
  ~Teacher, ~Student, ~Age,
   "0123",    1,       7,
   "0145",    1,       7,
   "0163",    1,       7,
   "0175",    2,       8,
   "0123",    2,       8,
   "0194",    2,       8,
   "0123",    3,       7,
   "0145",    3,       7
  )

df2 <- tibble::tribble(
  ~Teacher, ~Age.7, ~Age.8, ~Age.9,
     "0123",    1,       1,      1,
     "0145",    5,       7,      3,
     "0163",    4,       7,      1,
     "0175",    6,       8,      1,
     "0183",    3,       8,      1,
     "0194",    2,       8,      1,
     "0120",    3,       7,      4
  )

# Load necessary libs
library(dplyr, warn.conflicts = FALSE)
library(tidyr)

tidyr::pivot_longer() transforms df2 to a long format and dplyr::mutate() with gsub() and as.numeric() are used to shave of the residue from the variable names and convert to a dbl.

df2_long <-
  df2 %>%
  pivot_longer(Age.7:Age.9,
               names_to = "Age",
               values_to = "AgeStereotype") %>%
  mutate(Age = as.numeric(gsub("Age.", "", Age)))

dplyr::left_join() combines the datasets, only keeping those teachers, that have a row in df1 .

left_join(df1, df2_long)
#> Joining, by = c("Teacher", "Age")
#> # A tibble: 8 x 4
#>   Teacher Student   Age AgeStereotype
#>   <chr>     <dbl> <dbl>         <dbl>
#> 1 0123          1     7             1
#> 2 0145          1     7             5
#> 3 0163          1     7             4
#> 4 0175          2     8             8
#> 5 0123          2     8             1
#> 6 0194          2     8             8
#> 7 0123          3     7             1
#> 8 0145          3     7             5

Answer 4

Another base way:

merge(
  df1,
  data.frame(Teacher = df2$Teacher, 
             Age = gsub("[[:alpha:]]", "", stack(df2[,-1])[,2]),
             AgeStereotype = stack(df2[,-1])[,1]
  )
)

Output:

  Teacher Age Student AgeStereotype
1     123   7       1             1
2     123   7       3             1
3     123   8       2             1
4     145   7       1             5
5     145   7       3             5
6     163   7       1             4
7     175   8       2             8
8     194   8       2             8

This does change the original order though, and even if it is possible to amend this without additional packages, if it is important perhaps easiest is to just let dplyr do the join:

dplyr::left_join(
  df1,
  data.frame(Teacher = df2$Teacher, 
             Age = as.integer(gsub("[[:alpha:]]", "", stack(df2[,-1])[,2])),
             AgeStereotype = as.integer(stack(df2[,-1])[,1]), stringsAsFactors = FALSE
  )
)