Using dataframe column to specify dataframes for bind_rows in r

July 16, 2023

I have a dataframe (here, dfToc) containing a list of dataframe names. I would like to use it to specify other dataframes to concatenate via the tidyverse command bind_rows. The concatenation works if I explicitly name the other dataframes in the bind_rows command, but not if I use dfToc to specify the other dataframes. It seems elementary but I am stumped. A simplified example follows.

library(tidyverse)

df1 <- data.frame(Name = c("Jon", "Bill", "Maria", "Ben", "Tina"),
                  Score = c(23, 41, 32, 58, 26))

df2 <- data.frame(Name = c("Jon", "Bill", "Maria", "Ben", "Tina"),
                  Score = c(32, 14, 23, 85, 62))

dfToc <- data.frame(fn = c('df1' , 'df2'))
dfToc$fn
#> [1] "df1" "df2"

# explicitly naming the dataframes works
bind_rows(list(df1 , df2))
#>     Name Score
#> 1    Jon    23
#> 2   Bill    41
#> 3  Maria    32
#> 4    Ben    58
#> 5   Tina    26
#> 6    Jon    32
#> 7   Bill    14
#> 8  Maria    23
#> 9    Ben    85
#> 10  Tina    62

# using dfToc to specify the dataframes results in an error
bind_rows(list(dfToc$fn))
#> Error in `bind_rows()`:
#> ! Argument 1 must be a data frame or a named atomic vector.
#> Backtrace:
#>     ▆
#>  1. └─dplyr::bind_rows(list(dfToc$fn))
#>  2.   └─rlang::abort(glue("Argument {i} must be a data frame or a named atomic vector."))

Thanks for any help!

>Solution :

dfToc$fn is a character vector representing the names of your data frames. It is not a list of data frames, which is what bind_rows needs. To get multiple objects from your workspace into a list using the names of those objects, we can use mget:

bind_rows(mget(dfToc$fn))
#>     Name Score
#> 1    Jon    23
#> 2   Bill    41
#> 3  Maria    32
#> 4    Ben    58
#> 5   Tina    26
#> 6    Jon    32
#> 7   Bill    14
#> 8  Maria    23
#> 9    Ben    85
#> 10  Tina    62