Splitting a string without delimiters in R - regex

Question

I have a character vector in R, with each element containing a string - let's use this example:
my.files <- c("AWCallibration#NoneBino-3", "AWExperiment1#NoneBino-1", "AWExperiment2#NonemonL-2"
)
I would like to extract certain information out of these strings -
First, two uppercase alpha characters (in this case, always "AW")
Whether the trial was for calibration ("Callibration") or data collection - if it was the latter, which condition was used ("Experiment1" or "Experiment2")
Which sub-condition was used on this particular trial ("Bino" or "monL")
The repetition of the sub-condition ("1" or "2")
I first tried using strsplit, but this only appears to work for cases with regular delimiters such as "_". substring appeared to suit my needs better, but did not actually work due to the fact that splits don't occur in regular places ("Experiment1" is eleven elements long, "Callibration" is twelve).
I suspect that use of regular expressions may be the answer here, but I don't know how to account for the different lengths between the splits.

Answer 1

You can extract the information one by one:
first <- substr(my.files, 1, 2)
# [1] "AW" "AW" "AW"
second <- sub("^..(.*)#.*", "\\1", my.files)
# [1] "Callibration" "Experiment1" "Experiment2"
third <- sub("^.*#None(.*)-\\d+$", "\\1", my.files)
# [1] "Bino" "Bino" "monL"
fourth <- sub(".*-(\\d+)$", "\\1", my.files)
# [1] "3" "1" "2"
All in one command:
strsplit(my.files, "(?<=^..)(?=[A-Z])|#None|-", perl = TRUE)
# [[1]]
# [1] "AW" "Callibration" "Bino" "3"
#
# [[2]]
# [1] "AW" "Experiment1" "Bino" "1"
#
# [[3]]
# [1] "AW" "Experiment2" "monL" "2"

Answer 2

Here are a few different solutions:
gsubfn::strapplyc Try this:
library(gsubfn)
pat <- "(..)(.*)#None(.*)-(.*)"
strapplyc(my.files, pat, simplify = rbind)
which gives:
[,1] [,2] [,3] [,4]
[1,] "AW" "Callibration" "Bino" "3"
[2,] "AW" "Experiment1" "Bino" "1"
[3,] "AW" "Experiment2" "monL" "2"
Note that in the development version of the gsubfn package there is a read.pattern command which could use the above pat like this: read.pattern(text = my.files, pattern = pat, as.is = TRUE) .
sub/strsplit Here is an alternate solution. It inserts a minus after the second character and then splits each strip by minus or #None:
my.files2 <- sub("(..)", "\\1-", my.files)
do.call(rbind, strsplit(my.files2, "-|#None"))
which gives:
[,1] [,2] [,3] [,4]
[1,] "AW" "Callibration" "Bino" "3"
[2,] "AW" "Experiment1" "Bino" "1"
[3,] "AW" "Experiment2" "monL" "2"
gsub/read.table Here we use gsub to insert a minus after the first two characters and also we replace #None with minus. Then we use read.table with a sep of minus to read it in:
withMinus <- gsub("^(..)|#None", "\\1-", my.files)
read.table(text = withMinus, sep = "-", as.is = TRUE)
V1 V2 V3 V4
1 AW Callibration Bino 3
2 AW Experiment1 Bino 1
3 AW Experiment2 monL 2
REVISION:
A correction and a second solution.
Third solution.