Tidyverse: Multiply numeric columns from two data.frames-CodePudding

I have a data.frame of Quantities and other of Weights and want to multiply Quantities with their corresponding Weights. Can the task be accomplished by selecting only numeric columns? However, I want to keep the other columns in the final output too. Struggling how to get it more efficiently.

library(tidyverse)

df1 <-
  tibble(
    "A" = paste0("A", 1:10)
  , "Q" = 1:10
    )

df1
#> # A tibble: 10 x 2
#>    A         Q
#>    <chr> <int>
#>  1 A1        1
#>  2 A2        2
#>  3 A3        3
#>  4 A4        4
#>  5 A5        5
#>  6 A6        6
#>  7 A7        7
#>  8 A8        8
#>  9 A9        9
#> 10 A10      10

df2 <-
  tibble(
    "A" = paste0("A", 1:10)
  , "W" = 11:20
    )

df2
#> # A tibble: 10 x 2
#>    A         W
#>    <chr> <int>
#>  1 A1       11
#>  2 A2       12
#>  3 A3       13
#>  4 A4       14
#>  5 A5       15
#>  6 A6       16
#>  7 A7       17
#>  8 A8       18
#>  9 A9       19
#> 10 A10      20

select(df1, -A) * select(df2, -A)
#>      Q
#> 1   11
#> 2   24
#> 3   39
#> 4   56
#> 5   75
#> 6   96
#> 7  119
#> 8  144
#> 9  171
#> 10 200

df1*df2
#> Error in FUN(left, right): non-numeric argument to binary operator

Edited

My real data sets has many columns, so looking for a more generic solution.

CodePudding user response：

library(tidyverse)

df1 <-
  tibble(
    "A" = paste0("A", 1:10)
    , "Q" = 1:10
  )

df2 <-
  tibble(
    "A" = paste0("A", 1:10)
    , "W" = 11:20
  )

df1 |>
  mutate(Q*df2$W)
#> # A tibble: 10 x 3
#>    A         Q `Q * df2$W`
#>    <chr> <int>       <int>
#>  1 A1        1          11
#>  2 A2        2          24
#>  3 A3        3          39
#>  4 A4        4          56
#>  5 A5        5          75
#>  6 A6        6          96
#>  7 A7        7         119
#>  8 A8        8         144
#>  9 A9        9         171
#> 10 A10      10         200

^{Created on 2022-02-25 by the reprex package (v2.0.1)}

If df1$A == df2$A then you can use left_join and mutate as well.

CodePudding user response：

Another possible solution, using inner_join:

library(tidyverse)

df1 <-
  tibble(
    "A" = paste0("A", 1:10)
    , "Q" = 1:10
  )

df2 <-
  tibble(
    "A" = paste0("A", 1:10)
    , "W" = 11:20
  )

df1  %>% 
  inner_join(df2) %>% 
  mutate(T = Q*W)
#> Joining, by = "A"
#> # A tibble: 10 × 4
#>    A         Q     W     T
#>    <chr> <int> <int> <int>
#>  1 A1        1    11    11
#>  2 A2        2    12    24
#>  3 A3        3    13    39
#>  4 A4        4    14    56
#>  5 A5        5    15    75
#>  6 A6        6    16    96
#>  7 A7        7    17   119
#>  8 A8        8    18   144
#>  9 A9        9    19   171
#> 10 A10      10    20   200

CodePudding user response：

Using base R:

data.frame(A = df1$A, variable = df1$Q * df2$W)

     A variable
1   A1       11
2   A2       24
3   A3       39
4   A4       56
5   A5       75
6   A6       96
7   A7      119
8   A8      144
9   A9      171
10 A10      200

Using dplyr:

library(dplyr)
mutate(df1, variable = Q * df2$W) %>% select(-Q) # select(-Q) removes the old Q variable

# A tibble: 10 x 2
   A     variable
   <chr>    <int>
 1 A1          11
 2 A2          24
 3 A3          39
 4 A4          56
 5 A5          75
 6 A6          96
 7 A7         119
 8 A8         144
 9 A9         171
10 A10        200

In case you want to keep all the columns in the final product, you can do:

library(dplyr)
left_join(df1, df2, by = "A") %>% 
        mutate(variable = Q * W)

# A tibble: 10 x 4
   A         Q     W variable
   <chr> <int> <int>    <int>
 1 A1        1    11       11
 2 A2        2    12       24
 3 A3        3    13       39
 4 A4        4    14       56
 5 A5        5    15       75
 6 A6        6    16       96
 7 A7        7    17      119
 8 A8        8    18      144
 9 A9        9    19      171
10 A10      10    20      200