CODEWITHSUNDEEP
r
A Hud
A Hud

Reputation: 95

Multiple Condition Keyword Search

I’m trying to a key word search based on a couple of conditions. My code below creates some example data and tries to work this through. It doesn’t like the ifelse, I believe since GREP doesn’t actually return a true/false, but instead the actual data. I’m thinking I’d have to create a subset, then merge it back in? However, I can’t seem to put more than 1 condition in the subset with a GREP.

Any help is much appreciated!

bus_name <- c('jacks wok shop'
               ,'jacks ski shop'
               ,'jacks wokshop'
               ,'jacks bakery'
               ,'jacks business'
               ,'jims Brewery'
               ,'jims Wok Brewery'
)
type <- c('restaurant'
             ,'restaurant'
             ,'restaurant'
             ,'restaurant'
             ,'office'
             ,'store'
             ,'building'
)

mydata <- data.frame(bus_name, type)

mydata$bus_name <- as.character(mydata$bus_name)
mydata$type <- as.character(mydata$type)


mydata$flag <- ifelse(mydata$type == "restaurant" & mydata[grep("WOK",toupper(mydata$bus_name)),],"Wok",
                      ifelse(mydata$type == "office" & mydata[grep("BUSINESS",toupper(mydata$bus_name)),],"Business","0"))

The ideal output would be the following: 

bus_name|type|flag  
jacks wok shop|restaurant|Wok  
jacks ski shop|restaurant  
jacks wokshop|restaurant|Wok  
jacks bakery|restaurant  
jacks business|office|Business  
jims Brewery|store  
jims Wok Brewery|building|Wok

Upvotes: 1

Views: 152

Answers (2)

Pierre Lapointe
Pierre Lapointe

Reputation: 16277

You have to use grepl to get a logical vector of TRUE/FALSE. You could use case_when from dplyr. It's easier to read than nested ifelse.

library(dplyr)
mydata %>%
mutate(flag=case_when(
 .$type == "restaurant" & grepl("WOK",toupper(.$bus_name))  ~ "Wok",
 .$type == "office" & grepl("BUSINESS",toupper(.$bus_name)) ~ "Business",
                                                       TRUE ~ "0"
))

          bus_name       type     flag
1   jacks wok shop restaurant      Wok
2   jacks ski shop restaurant        0
3    jacks wokshop restaurant      Wok
4     jacks bakery restaurant        0
5   jacks business     office Business
6     jims Brewery      store        0
7 jims Wok Brewery   building        0

Upvotes: 2

Andrew Gustar
Andrew Gustar

Reputation: 18425

grepl is the function to use if you want logical values. You need something like this

mydata$flag <- ifelse(mydata$type == "restaurant" & grepl("WOK",toupper(mydata$bus_name)),
                              "Wok",
                      ifelse(mydata$type == "office" & grepl("BUSINESS",toupper(mydata$bus_name)),
                              "Business",""))

Upvotes: 1

Related Questions