根据选定的变量/列名称过滤和设置R数据帧

Question

I am trying to subset a large data set with many variables/columns names, say ax1, ax2, ax3, ax4, ax5, ...,ax20, bx1...bx20...zx1...zx20. For example, suppose the subset data I want to obtain are on variables ax3, ax5, ax11, ax19,..., bx3, bx5, cx11, cx19,...,zx3, zx5, zx11, zx19.

I have tried the following code in R but it is becoming very lengthy and cumbersome.

setwd("")
abc<- read.table("abc.txt", header=TRUE)
new.abc<-data.frame(abc$ax3,abc$ax5,abc$ax5,abc$ax11,abc$ax19,  
abc$bx3,abc$bx5,abc$bx5,abc$bx11,abc$bx19)

The code is becoming longer as I need to continue with cx3, cx5, cx11, cx19,...,zx3, xz5, zx11, zx19. I am looking for an alternative approach that can avoid this lengthy coding. Your help is greatly appreciated.

Answer 1

You could create columns programmatically. If they follow the same structure as mentioned in the question, we can do

cols <- c(outer(paste0(letters, "x"), c(3, 5, 11, 19), paste0))
cols
#[1] "ax3"  "bx3"  "cx3"  "dx3"  "ex3"  "fx3"  "gx3"  "hx3"  "ix3"  "jx3"  "kx3"...

and then use it to subset the dataframe

new.abc[, cols]

If we also want to preserve column order, we can use gtools::mixedsort

new.abc[, gtools::mixedsort(cols)]