在 R 中的多个观察中使用 function “substr”

Question

I have a dataset of 100 observations and one variable, with each observations being a string of integers. I would like to substract integers from each observation and create a new data frame with the same number of observations but with each string divided into several variables.

Basically I would like to go from this:

Variable 1
1234567
1234567
1234567

To this

Variable 1   Variable 2   Variable 3

123             456         7
123             456         7
123             456         7

I have tried using the function substr to do it, but while it works correctly when I use it with a subset of only 1 observation, it does not appear to work when I use it over the whole dataset. Any ideas about how could I use substr here, or if there is a better alternative for it?

Answer 1

Assuming your dataset is called df , with the column you want to split called Var1 :

tidyr::separate(df, Var1, into = c("Var1", "Var2", "Var3"), sep = c(3, 6, 7))

#   Var1 Var2 Var3
# 1  123  456    7
# 2  123  456    7
# 3  123  456    7

The sep argument takes a vector indicating the positions at which to split.

Answer 2

In base R , we can also use sub to create a delimiter at the specified location and then with read.csv can read it

read.csv(text = sub("^(...)(...)(.)$", "\\1,\\2,\\3", 
     df1$Variable1), header = FALSE, col.names = paste0("Variable", 1:3))
#  Variable1 Variable2 Variable3
#1       123       456         7
#2       123       456         7
#3       123       456         7

---

Or as @markus mentioned in the comments, read.fwf can be used along with textConnection

read.fwf(textConnection(paste(df1$Variable1, collapse="\n")),
         widths = c(3, 3, 1), as.is = TRUE)
#  V1  V2 V3
#1 123 456  7
#2 123 456  7
#3 123 456  7

data

df1 <- structure(list(Variable1 = c(1234567L, 1234567L, 1234567L)), 
      class = "data.frame", row.names = c(NA, 
-3L))     

Answer 3

You could use substr with mapply .

dat <- cbind(dat, mapply(function(...) as.double(substr(...)), list(dat$v1), 
                         c(1, 4, 7), c(3, 6, 7)))
dat
#        v1          v2   1   2 3
# 1 1234567 -0.60679296 123 456 7
# 2 1234567 -0.06347641 123 456 7
# 3 1234567 -0.58993170 123 456 7
# 4 1234567 -0.71293088 123 456 7
# 5 1234567 -0.28107903 123 456 7

---

Data

dat <- data.frame(v1=1234567, v2=rnorm(5))