CODEWITHSUNDEEP
r
Gecko
Gecko

Reputation: 390

How to generate multiple data frames in a loop

I want to generate multiple data frames (df1, df2 etc.) based on one big data frame (df0). Each new data frame should consist of some mix of columns from df0.

df0 <- data.frame(v0=c(0, 0), v1=c(3, 4), v2=c(5, 6), v3=c(7, 8))

for(i in 1:3) {
   secondcol <- colnames(df0[,..i])                  # I get an error here
   dfX = subset(df0, select = c("v0", secondcol))    # dfX should be df & i
}

# The for loop should replicate the following three comands:
df1 = subset(df0, select = c("v0", "v1"))
df2 = subset(df0, select = c("v0", "v2"))
df3 = subset(df0, select = c("v0", "v3"))

Upvotes: 0

Views: 72

Answers (2)

Sonny
Sonny

Reputation: 3183

Many ways to do that. To make your code work as below:

library(dplyr)
for(i in 1:3) {
  secondcol <- colnames(df0)[(i+1)]
  dat <- select(df0, c("v0", secondcol)) 
  assign(paste0("df", i), dat)
  rm(dat)
}

@Rich solution is much better to manage. In case, you still want them as dataframes in your environment, you can use:

i <- 1:3
list2env(setNames(lapply(i+1, function(j) df0[c(1, j)]), paste0("df", i)), envir = .GlobalEnv)

Upvotes: 2

Rich Scriven
Rich Scriven

Reputation: 99321

I would recommend using a list to keep the resulting data frames organized. Here is an approach using lapply().

i <- 1:3
setNames(lapply(i+1, function(j) df0[c(1, j)]), paste0("df", i))
# $df1
#   v0 v1
# 1  0  3
# 2  0  4
#
# $df2
#   v0 v2
# 1  0  5
# 2  0  6
#
# $df3
#   v0 v3
# 1  0  7
# 2  0  8

Upvotes: 4

Related Questions