CODEWITHSUNDEEP
rstringr
Maria Faleeva
Maria Faleeva

Reputation: 57

How to extract a word from a column to their own separate column in r

For example I have a column where the characters go something like this

"Object=house colour=blue size=big", "Object=roof colour=red size=small", "Object=window colour=green size=medium"

I want to extract just the word after colour and create a new column. so it would look like this

"blue", "red", "green"

I have started trying to do this with str_extract but am very lost on how to specify things. So far I have

colour<- str_extract(string = df, pattern = "(?<=colour= ).*(?=\\,)")

How would I go about solving this?

Upvotes: 1

Views: 57

Answers (4)

hello_friend
hello_friend

Reputation: 5788

Base R solution:

df <- c(
  "Object=house colour=blue size=big", 
  "Object=roof colour=red size=small", 
  "Object=window colour=green size=medium"
)

res <- data.frame(
  object_info = df,
  object_colour = gsub(
    ".*\\s+colour\\=(\\S+).*",
    "\\1",
    df
  ),
  row.names = NULL,
  stringsAsFactors = FALSE
)

res

Upvotes: 0

Onyambu
Onyambu

Reputation: 79218

You can also convert to dcf, and turn this to a data.frame:

read.dcf(textConnection(paste(chartr("= ", ":\n", text), collapse = "\n\n")), all =TRUE)
  Object colour   size
1  house   blue    big
2   roof    red  small
3 window  green medium

Then you can select the column you want

Upvotes: 1

akrun
akrun

Reputation: 887108

There was no space after the = and also, we can use \\S+ to specify one or more non-white space

library(stringr)
str_extract(string = df, pattern = "(?<=colour=)\\S+")
[1] "blue"  "red"   "green"

data

df <- c("Object=house colour=blue size=big", "Object=roof colour=red size=small", 
"Object=window colour=green size=medium")

Upvotes: 2

PaulS
PaulS

Reputation: 25323

A possible solution:

library(tidyverse)

c("Object=house colour=blue size=big", "Object=roof colour=red size=small", "Object=window colour=green size=medium") %>% 
  str_extract("(?<=colour\\=)\\S+")

#> [1] "blue"  "red"   "green"

Upvotes: 1

Related Questions