Pattern matching character strings when everything but the last two elements of the character are th-CodePudding

I have the following vector.

column_names <- c("6Li", "7Li", "10B", "11B", "7Li.1",
                  "205Pb", "206Pb", "207Pb", "238U",
                  "206Pb.1", "238U.1")

Notice that some of the values are just duplicates with a ".1" stuck at the end. I want to index out all of these character strings along with their corresponding character strings that match such that only the following are returned.

#[1] "7Li"     "7Li.1"   "206Pb"   "238U"    "206Pb.1" "238U.1"

Assume you don't know the index positions and so you cannot simply index these values out as follows column_names[c(2,5,7,9,10,11)]. How can I use pattern matching to extract these values?

CodePudding user response：

There is likely a more elegant solution, but in base R you cold try a combination of grep/gsub and paste:

idx <- grep(paste(gsub("\\.1", "", column_names[grep("\\.1", column_names)]), collapse = "|"), column_names)
# [1]  2  5  7  9 10 11

column_names[idx]
# [1] "7Li"     "7Li.1"   "206Pb"   "238U"    "206Pb.1" "238U.1"

CodePudding user response：

Using gsub() and duplicated() to find values with repeated stems:

column_stems <- gsub("\\.1", "", column_names)

dup_idx <- duplicated(column_stems) | duplicated(column_stems, fromLast = TRUE)

column_names[dup_idx]
# "7Li"     "7Li.1"   "206Pb"   "238U"    "206Pb.1" "238U.1"

To also find instances ending with .2, .3, etc., use "\\.\\d " instead of "\\.1" in gsub().

CodePudding user response：

You could use stringr:

library(stringr)

idx <- str_extract(column_names, ".*(?=\\.1)")

column_names[str_detect(column_names, paste(idx[!is.na(idx)], collapse = "|"))]

This returns

#> [1] "7Li"     "7Li.1"   "206Pb"   "238U"    "206Pb.1" "238U.1"