How to extract one element of text from a column in R? - r

Question

I'm working with a data frame that contains the locations of where people got tested for COVID. There is not standardization of formatting of the ordering facility (the place that ordered the test). My data frame look something like this:
TestingLocation <- data.frame(TestingLocation= c("New York Hospital One", "Chicago Clinic Two", "Nursing Home Name One",
"Los Angeles University_Testing_Site", "Test-Site-in-BOSTON-MA"))
I have a list of the cities where someone could get tested.
Cities <- data.frame(PossibleTestCities=c("Los Angeles", "Chicago", "New York", "Miami", "Boston", "Austin", "Santa Fe"))
Is there a way to use the Cities frame I have to extract the city and put it into a new column. Additionally, if no city appears, to put "Unknown" or something along those lines? Ideally, my frame would look like this:
DesiredFrame <- data.frame(TestingLocation= c("New York Hospital One", "Chicago Clinic Two", "Nursing Home Name One",
"Los Angeles University_Testing_Site", "Test-Site-in-BOSTON-MA"),
TestCity= c("New York", "Chicago", "Unknown", "Los Angeles", "Boston"))
Thank you!

Answer 1

Does this work:
library(dplyr)
library(stringr)
TestingLocation %>% mutate(TestCity = str_to_title(str_extract(toupper(TestingLocation), toupper(str_c(Cities$PossibleTestCities, collapse = '|'))))) %>%
mutate(TestCity = replace_na(TestCity, 'Unknown'))
TestingLocation TestCity
1 New York Hospital One New York
2 Chicago Clinic Two Chicago
3 Nursing Home Name One Unknown
4 Los Angeles University_Testing_Site Los Angeles
5 Test-Site-in-BOSTON-MA Boston

Answer 2

This doesn't look pretty but it works:
TestingLocation$TestCity <- sub("(^[a-z]+.*$)", NA, sub(paste0(".*(",
paste(tolower(Cities$PossibleTestCities), collapse = "|"),").*"),
"\\U\\1", tolower(TestingLocation$TestingLocation), perl = T))
There's a number of operations involved. There are two sub operations one nested in the other. The first is to replace the (lower-case) TestingLocation$TestingLocations with the matching (lower-case) Cities$PossibleTestCities and set the replacements to upper-case, while the second is to set the values that did not find a match and that hence remained lower-case to NA.
Instead of using a compact but hard-to parse single piece of code you can achieve the substitutions step-by-step:
# 1. define pattern with alternatives:
mypattern <- paste0(".*(", paste(tolower(Cities$PossibleTestCities), collapse = "|"),").*")
# 2. perform first substitution to set matches to City names:
TestingLocation$TestCity <- sub(mypattern, "\\U\\1", tolower(TestingLocation$TestingLocation), perl = T)
# 3. perform second substitution to set non-match to NA:
TestingLocation$TestCity <- sub("(^[a-z]+.*$)", NA, TestingLocation$TestCity)
Result:
TestingLocation
TestingLocation TestCity
1 New York Hospital One NEW YORK
2 Chicago Clinic Two CHICAGO
3 Nursing Home Name One <NA>
4 Los Angeles University_Testing_Site LOS ANGELES
5 Test-Site-in-BOSTON-MA BOSTON