Finding the Matching Characters after Pattern in R DataFrame - r

Question

I am fairly new to string manipulation, and I am stuck on a problem regarding string and character data in an R dataframe. I am attempting to extract numeric values from a long string after a pattern and then store the result as a new column in my dataframe. I have a fairly large dataset, and I am attempting to get out some useful information stored in a column called "notes".
For instance, the strings I am interested in always follow this pattern (there is nothing significant about the tasks):
df$notes[1] <- "On 5 June, some people walked down the street in this area. [size=around 5]"
df$notes[2] <- "On 6 June, some people rode bikes down the street in this area. [size= nearly 4]"
df$notes[3] <- "On 7 June, some people walked into a grocery store in this area. [size= about 100]"
In some columns, we do not get a numeric value, and that is a problem I can deal with after I get a solution to this one. Those rows follow something similar to this:
df$notes[4] <- "On 10 July, an hundreds of people drank water from this fountain [size=hundreds]"
df$notes[5] <- "on 15 August, an unreported amount of people drove their cars down the street. [size= no report]"
I am trying to extract the entire match after "size= (some quantifier)", and store the value into an appended column of my dataframe.
Eventually, I need to write a loop that goes through this column (call it "notes") in my dataframe, and storing the values "5, 4, 100" into a new column (call it "est_size").
Ideally, my new column will look like this:
df$est_size[1] <- "around 5"
df$est_size[2] <- "nearly 4"
df$est_size[3] <- "about 100"
df$est_size[4] <- "hundreds"
df$est_size[5] <- "no report"
Code that I have tried / stuck on:
stringr::str_extract(notes[1], \w[size=]\d"
but all I get back is "size=" and not the value after
Thank you in advance for helping!

Answer 1

We can use a regex lookaround to match one or more characters that are not a closing square bracket ] after the size=
library(dplyr)
library(stringr)
df <- df %>%
mutate(est_size = trimws(str_extract(notes, '(?<=size=)[^\\]]+')))
-output
df #notes est_size
#1 On 5 June, some people walked down the street in this area. [size=around 5] around 5
#2 On 6 June, some people rode bikes down the street in this area. [size= nearly 4] nearly 4
#3 On 7 June, some people walked into a grocery store in this area. [size= about 100] about 100
#4 On 10 July, an hundreds of people drank water from this fountain [size=hundreds] hundreds
#5 on 15 August, an unreported amount of people drove their cars down the street. [size= no report] no report
data
df <- structure(list(notes = c("On 5 June, some people walked down the street in this area. [size=around 5]",
"On 6 June, some people rode bikes down the street in this area. [size= nearly 4]",
"On 7 June, some people walked into a grocery store in this area. [size= about 100]",
"On 10 July, an hundreds of people drank water from this fountain [size=hundreds]",
"on 15 August, an unreported amount of people drove their cars down the street. [size= no report]"
)), class = "data.frame", row.names = c(NA, -5L))

Answer 2

Using str_extract:
library(stringr)
trimws(str_extract(df$notes, "(?<=size=)[\\w\\s]+"))
[1] "around 5" "nearly 4" "about 100" "hundreds" "no report"
Here, we use positive lookbehind (?<=...) to assert an accompanying pattern for what we want to extract: we want to extract the alphanumeric string(s) that follow after size=so we put size=into the lookbehind expression and extract whatever alphanumeric chars (\\w) and whitespace chars (\\s) (but not special chars such as ]!) come after it.