CODEWITHSUNDEEP
regexrdplyr
emehex
emehex

Reputation: 10538

gsub() an entire column over another column

I have a df:

df <- data.frame(
    x=c("ABC Inc", "DCV", "FGZ", "JH7 j11"),
    y=c("ABC - fasjdlkjs", "DCV . (INC) .. kdhkfhksf", "FGZ / qiuwy72gs", "JH7 j11 dhd"),
    target=c("fasjdlkjs", "inc kdhkfhksf", "qiuwy gs", "dhd")
)

Where x is a close, but not exact subset of y

I want to gsub() everything in x to "" (blank) in y, while also removing numbers/punctuation.

My desired output is stored in target

I thought this would have worked, but it didn't:

df <- mutate(target = gsub(pattern=x, replacement="", y))

EDIT:

Sort of: Y - X = Target

Upvotes: 1

Views: 106

Answers (1)

Rorschach
Rorschach

Reputation: 32426

This (now - thanks @Frank) converts case tolower. Below, s builds the string to test against from the x column by splitting the x string by spaces

df$res <- mapply(function(a, b) {
    s <- paste(c(unlist(strsplit(as.character(a)," ")), "[[:punct:]]"), collapse="|")
    tolower(gsub("[[:digit:]]+", " ", gsub(s, "", b)))
  }, df$x, df$y)

df
#         x                        y        target              res
# 1 ABC Inc          ABC - fasjdlkjs     fasjdlkjs        fasjdlkjs
# 2     DCV DCV . (INC) .. kdhkfhksf inc kdhkfhksf   inc  kdhkfhksf
# 3     FGZ          FGZ / qiuwy72gs      qiuwy gs         qiuwy gs
# 4 JH7 j11              JH7 j11 dhd           dhd              dhd

Upvotes: 3

Related Questions