Extract & combine multiple substrings using multiple patterns from some but not all strings contained in list & return to list in R

问题

I'd like to find an elegant and easily manipulable way to:

extract multiple substrings from some, but not all, strings that are contained as elements of a list (each list element consists of just one long string)
replace the respective original long string with these multiple substrings
collapse the substrings in each list element into 1 string
return a list of same length containing the replacement substrings and the untouched long strings as appropriate.

This question is a follow-on (though different) from my earlier question: replace strings of some list elements with substring. Note, I don't want to run the regex patterns over all list elements, only those elements to which the regex applies.

I know the end result can be delivered by str_replace or sub by matching the entire strings to be changed and returning the text captured by capturing groups, as follows:

library(stringr)
myList <- as.list(c("OneTwoThreeFourFive", "mnopqrstuvwxyz", "ghijklmnopqrs", "TwentyTwoFortyFourSixty"))
fileNames <- c("AB1997R.txt", "BG2000S.txt", "MN1999R.txt", "DC1997S.txt")
names(myList) <- fileNames
is1997 <- str_detect(names(myList), "1997")

regexp <- ".*(Two).*(Four).*"
myListNew2 <- myList
myListNew2[is1997] <- lapply(myList[is1997], function(i) str_replace(i, regexp, "\\1££\\2"))

## This does return what I want:
myListNew2
$AB1997R.txt
[1] "Two££Four"

$BG2000S.txt
[1] "mnopqrstuvwxyz"

$MN1999R.txt
[1] "ghijklmnopqrs"

$DC1997S.txt
[1] "Two££Four"

But I would prefer do it without having to match the entire original text (because, e.g., of time required for matching very long texts; of complexity of multiple regex patterns & difficulty of knitting them together so they match entire strings successfully). I would like to use separate regex patterns to extract the substrings and then replace the original string with these extracts. I came up with the following, which works. But surely there is an easier, better way! llply?

patternA <- "Two"
patternB <- "Four"
x <- myList[is1997]
x2 <- unlist(x)
stringA <- str_extract (x2, patternA)
stringB <- str_extract (x2, patternB)
x3 <- mapply(FUN=c, stringA, stringB, SIMPLIFY=FALSE)
x4 <- lapply(x3, function(i) paste(i, collapse = "££"))
x5 <- relist(x4,x2)
myListNew1 <- replace(myList, is1997, x5)
myListNew1

$AB1997R.txt
[1] "Two££Four"

$BG2000S.txt
[1] "mnopqrstuvwxyz"

$MN1999R.txt
[1] "ghijklmnopqrs"

$DC1997S.txt
[1] "Two££Four"

回答1:

Something like this maybe, where I've extended the patterns you are looking for to show how it could become adaptable:

library(stringr)
patterns <- c("Two","Four","Three")
hits <- lapply(myList[is1997], function(x) {
  out <- sapply(patterns, str_extract, string=x)
  paste(out[!is.na(out)],collapse="££")
})
myList[is1997] <- hits

#[[1]]
#[1] "Two££Four££Three"
#
#[[2]]
#[1] "mnopqrstuvwxyz"
#
#[[3]]
#[1] "ghijklmnopqrs"
#
#[[4]]
#[1] "Two££Four"

来源：https://stackoverflow.com/questions/30632886/extract-combine-multiple-substrings-using-multiple-patterns-from-some-but-not

标签

regex

string

text

plyr