Replace certain spaces to tabs - delimiters - regex

Question

I have one column data.frame where some spaces should be delimiters some just a space.
#input data
dat <- data.frame(x=c("A 2 2 textA1 textA2 Z1",
"B 4 1 textX1 textX2 textX3 Z2",
"C 3 5 textA1 Z3"))
# x
# 1 A 2 2 textA1 textA2 Z1
# 2 B 4 1 textX1 textX2 textX3 Z2
# 3 C 3 5 textA1 Z3
Need to convert it to 5 column data.frame:
#expected output
output <- read.table(text="
A 2 2 textA1 textA2 Z1
B 4 1 textX1 textX2 textX3 Z2
C 3 5 textA1 Z3",sep="\t")
# V1 V2 V3 V4 V5
# 1 A 2 2 textA1 textA2 Z1
# 2 B 4 1 textX1 textX2 textX3 Z2
# 3 C 3 5 textA1 Z3
Essentially, need to change 1st, 2nd, 3rd, and the last space to a tab (or any other delimiter if it makes it easier to code).
Playing with regex is not giving anything useful yet...
Note1: In real data I have to replace 1st, 2nd, 3rd,...,19th and the last spaces to tabs.
Note2: There is no pattern in V4, text can be anything.
Note3: Last column is one word text with variable length.

Answer 1

Try
v1 <- gsub("^([^ ]+)\\s+([^ ]+)\\s+([^ ]+)\\s+", '\\1,\\2,\\3,', dat$x)
read.table(text=sub(' +(?=[^ ]+$)', ',', v1, perl=TRUE), sep=",")
# V1 V2 V3 V4 V5
#1 A 2 2 textA1 textA2 Z1
#2 B 4 1 textX1 textX2 textX3 Z2
#3 C 3 5 textA1 Z3
Or an option inspired from #Tensibai's post
n <- 3
fpat <- function(n){
paste0('^((?:\\w+ ){', n,'})([\\w ]+)\\s+(\\w+)$')
}
read.table(text=gsub(fpat(n), "\\1'\\2' \\3", dat$x, perl=TRUE))
# V1 V2 V3 V4 V5
#1 A 2 2 textA1 textA2 Z1
#2 B 4 1 textX1 textX2 textX3 Z2
#3 C 3 5 textA1 Z3
For more columns,
n <- 19
v1 <- "A 24 34343 212 zea4 2323 12343 111 dsds 134d 153xd 153xe 153de 153dd dd dees eese tees3 zee2 2353 23335 23353 ddfe 3133"
read.table(text=gsub(fpat(n), "\\1'\\2' \\3", v1, perl=TRUE), sep='')
# V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12 V13 V14 V15
#1 A 24 34343 212 zea4 2323 12343 111 dsds 134d 153xd 153xe 153de 153dd dd
# V16 V17 V18 V19 V20 V21
#1 dees eese tees3 zee2 2353 23335 23353 ddfe 3133

Answer 2

With a variable number of columns:
library(stringr)
cols <- 3
m <- str_match(dat$x, paste0("((?:\\w+ ){" , cols , "})([\\w ]+) (\\w+)"))
t <- paste0(gsub(" ", "\t", m[,2]), m[,3], "\t", m[,4])
> read.table(text=t,sep="\t")
V1 V2 V3 V4 V5
1 A 2 2 textA1 textA2 Z1
2 B 4 1 textX1 textX2 textX3 Z2
3 C 3 5 textA1 Z3
Change the number of columns to tell how many you wish before.
For the regex:
((?:\\w+ ){3}) Capture the 3 repetitions {3} of the non capturing group (?:\w+ ) which matche at least one alphanumeric character w+ followed by a space
([\\w ]+) (\w+) capture the free text from alphanumeric char or space [\w ]+ followed by a space and the capture the last word with \w+
Once that done, paste the 3 parts returned by str_match taking care of replacing the spaces in the first group m[,2] by tabs.
m[,1] is the whole match so it's unused here.
Old answer:
A basic one matching based on a fixed number of fields:
> read.table(text=gsub("(\\w+) (\\w+) (\\w+) ([\\w ]+) (\\w+)$","\\1\t\\2\t\\3\t\\4\t\\5",dat$x,perl=TRUE),sep="\t")
V1 V2 V3 V4 V5
1 A 2 2 textA1 textA2 Z1
2 B 4 1 textX1 textX2 textX3 Z2
3 C 3 5 textA1 Z3
Add as many (\w+) you wish before, and increase the number of \1 (back references)

Answer 3

Here can be one twisted way to go that will work whatever the number of "words" you have (and that works on your data); it's based on the number of alphanum characters in your "words" compared to the number of alphanum characters in the other fields:
res <- gsub("\\w{3,}\\K\\t(?=\\w{3,})", " ", gsub(" ", "\t", dat$x), perl=T)
res
# [1] "A\t2\t2\ttextA1 textA2\tZ1" "B\t4\t1\ttextX1 textX2 textX3\tZ2" "C\t3\t5\ttextA1\tZ3"
read.table(text=res, sep="\t")
# V1 V2 V3 V4 V5
#1 A 2 2 textA1 textA2 Z1
#2 B 4 1 textX1 textX2 textX3 Z2
#3 C 3 5 textA1 Z3
EDIT: A completely different way to go, only based on the number of the spaces k you need to replace before the last one:
k <- 3 # in your example
res <- sapply(as.character(dat$x),
function(x, k){
pos_sp <- gregexpr(" ", x)[[1]]
x <- strsplit(x, "")[[1]]
if (length(pos_sp) > k+1) pos_sp <- pos_sp[c(1:k, length(pos_sp))]
x[pos_sp] <- "\t"
x <- paste(x, collapse="")
}, k=k)
read.table(text=res, sep="\t")
# V1 V2 V3 V4 V5
# 1 A 2 2 textA1 textA2 Z1
# 2 B 4 1 textX1 textX2 textX3 Z2
# 3 C 3 5 textA1 Z3