2

I am able to read in a subset of columns defined in cols_only like so:

x <- read_csv(filePath, col_types=cols_only(colA=col_character())
x <- read_csv(filePath, col_types=cols_only(colA='c'))

Both work fine, but I tried to create a list to pass into cols_only and splice it, like I do in many Tidyverse functions, but this one generates an error.

cols <- list(colA='c')
x <- read_csv(filePath, col_types=cols_only(!!!cols))

I tried this also as a named vector cols <- c(colA='c') but I get the error Error in !cols: invalid argument type.

Is this a limitation of the function? I saw this GitHub issue https://github.com/tidyverse/readr/issues/971 so was hoping this function could support this functionality.

Raesu
  • 310
  • 2
  • 15

1 Answers1

1

If the splicing is not working, use do.call

library(readr)
df1 <- read_csv(file.choose(), col_types = do.call(cols_only, cols))

Using a small reproducible example

cols1 <- list(`Service Charge`='c', `MONTHLY FEES` = 'c')
df1 <- read_csv(file.choose(), 
        col_types = do.call(cols_only, cols1))

-output

head(df1)
# A tibble: 6 x 2
#  `Service Charge`      `MONTHLY FEES`            
#  <chr>                 <chr>                     
#1 Customer Transfer Dr. INVESTMENT PURCHASE       
#2 Customer Transfer Dr. INVESTMENT PURCHASE       
#3 Miscellaneous Payment PAYPAL
akrun
  • 874,273
  • 37
  • 540
  • 662
  • Excellent, it works. Do you know why splicing doesn't work in this case? Curious about the implementation and what exactly is missing. – Raesu May 21 '21 at 19:35
  • @Raesu I am not sure about the reason. May be it is a bug or may be the readr package is included in the ambit of splicing . I expect it to work although the splicing works in most other cases. – akrun May 21 '21 at 19:37