0

I've got let's say 10 csv files with names like as

file_1_tail.csv 
file_2_tail.csv 
file_3_tail.csv  
...  
file_10_tail.csv

The only difference in name is in the number (1 to 10). Each has the same structure - 1000 rows and 100 columns.

I need to read them to the R, select specific columns and write as new file. My code for one file is below:

file_2_tail = read_csv("file_2_tail.csv")
file_2_tail_selected = file_2_tail[,c(1:7,30)])
write.csv2(file_2_tail_selected, file = "file_2_selected.csv")

And now I want to use loop to automate this for all ten files. 

for (i in 1:10){    
file_"i"_tail = read_csv("file_"i"_tail.csv")
file_"i"_tail_selected = file_"i"_tail[,c(1:7,30)]
write.csv2(file_"i"_tail_selected, file = "file_"i"_selected.csv")
}

And of course it doesn't work - i is not readable in this notation. How should I fix this?

anba
  • 543
  • 1
  • 4
  • 7

1 Answers1

1

You can't assign read_csv results to a string like that. Instead you can just store it in a temporary variable tmp

for (i in 1:10){    
  tmp <- read_csv(paste0("file_", i, "_tail.csv"))
  tmp <- tmp[, c(1:7,30)]
  write.csv2(tmp, file = paste0("file_", i, "_selected.csv"))
}

Btw this is probably a more efficient way to read multiple files

library(tidyverse)

filePattern <- "\\.csv$"
fileList <- list.files(path = ".", recursive = FALSE,
                       pattern = filePattern, full.names = TRUE)

result <- fileList %>%
  purrr::set_names(nm = (basename(.) %>% tools::file_path_sans_ext())) %>%
  purrr::map_df(read_csv, .id = "FileName") %>% 
  select(1:7, 30)
result
Tung
  • 26,371
  • 7
  • 91
  • 115
  • @anba: not a problem. Btw I updated my answer to include a more efficient way to read multiple files – Tung Aug 28 '18 at 15:01
  • @anba: btw, you can select the columns you want directly within `readr::read_csv` to make it more efficient. See this answer https://stackoverflow.com/a/50652089/ – Tung Aug 28 '18 at 15:12