I have this table

It contains columns that have a name and then a number in brackets, like you see: [1], [2], [3].. This is just an example with numbers from 1 to 3, but it can be bigger, like a 100. I would like to have a script who selects and put in a new df all the columns containing in the name [1], [2], [3]...[n]. I know that I can select with tidy verse with select(contains("..")) but since the numbers in the brackets can change in my initial table (they can be 1 to 3 or 1 to 100) I need something applicable to different cases. I was thinking of maybe generating a string from 1:n, but the problem is that the file I have always have the number in brackets []. And I can't just use contain with the numbers because unfortunately the names of the columns contains characters and numbers. This is the result I am looking for:

New table 1

New table 2

New table 3

CodePudding user response：

You can use data.table, in the following way

library(data.table)

setDT(df)

lapply(1:3, \(i)  {
  cols = (1:ncol(df))[grepl(paste0("\\[",i,"\\]"),names(df))]
  df[,.SD, .SDcols = cols]
})

Output:

[[1]]
   abc[1] abc[1] def[1] def[1]
1:    One    Two    One    Two
2:    One    Two    One    Two
3:    One    Two    One    Two

[[2]]
   abc[2] abc[2] def[2] def[2]
1:  Three    One  Three    One
2:  Three    One  Three    One
3:  Three    One  Three    One

[[3]]
   abc[3] abc[3] def[3] def[3]
1:    Two  Three    Two  Three
2:    Two  Three    Two  Three
3:    Two  Three    Two  Three

Input:

df = structure(list(`abc[1]` = c("One", "One", "One"), `abc[1]` = c("Two", 
"Two", "Two"), `abc[2]` = c("Three", "Three", "Three"), `abc[2]` = c("One", 
"One", "One"), `abc[3]` = c("Two", "Two", "Two"), `abc[3]` = c("Three", 
"Three", "Three"), `def[1]` = c("One", "One", "One"), `def[1]` = c("Two", 
"Two", "Two"), `def[2]` = c("Three", "Three", "Three"), `def[2]` = c("One", 
"One", "One"), `def[3]` = c("Two", "Two", "Two"), `def[3]` = c("Three", 
"Three", "Three")), row.names = c(NA, -3L), class = "data.frame")