Error about 4 columns even though am only asking for 3

[ggrothendieck]

If we run read_csv asking for 3 columns it gives an error if any of the files have more. I would have expected it to work by simply ignoring the columns that are not asked for.

library(readr)
packageVersion("readr")
## [1] ‘2.1.4’

fileNames <- c("test1.csv", "test2.csv")
cat("a,b,c\n1,2,3", file = fileNames[1])
cat("a,b,c,d\n11,12,13,14", file = fileNames[2])

read_csv(fileNames, col_select = 1:3, skip = 1, col_names = c("A", "B", "C"), id = "id")
## Error: Files must all have 3 columns:
## * File 2 has 4 columns

[hadley]

That's because the column specification has to be identical when writing multiple files. If you want to handle heterogenous files, we recommend using purrr or similar, e.g. https://r4ds.hadley.nz/iteration.html#heterogeneous-data.