Match and replace character in column in R - r

Question

I have been working on the following issue and cannot seem to solve it. Please give me your suggestions on how to solve it.
Let's say I have the following data frame.
NAICS 2017 NAICS 2012_1 NAICS 2012_2 NAICS 2012_3 NAICS 2012_4
2100 2111 0 0 0
9110 9119 5114 0 0
1113 5676 4875 2186 1153
6220 6225 1293 0 0
1115 3234 2163 0 0
7110 7873 0 0 0
1100 2679 8153 2114 1145
I want to essentially replace the NAICS 2017 column with the matching 2-digit NAICS if it exists within the other 4 NAICS columns.
So the code would determine if there is a 2 digit match (2100 matches 2111) then replace the 2 digit code with the four digit code. (2110 becomes 2111).
Here is how the final code would look.
NAICS 2017 NAICS 2012_1 NAICS 2012_2 NAICS 2012_3 NAICS 2012_4
2111 2111 0 0 0
9119 9119 5114 0 0
1153 5676 4875 2186 1153
6225 6225 1293 0 0
1115 3234 2163 0 0
7110 7873 0 0 0
1145 2679 8153 2114 1145
Optional Addition: Only change the NAICS code if the NAICS 2017 column is a 2 or 3 digit code (i.e. 2100 or 2110).
Could this be done with a grepl or gsub code?
If you would like a full data set please don't hesitate to ask.

Answer 1

Try this, the update column is the result:
library(dplyr)
it <- data.table::fread(
" NAICS_2017 NAICS 2012_1 NAICS 2012_2 NAICS 2012_3 NAICS 2012_4
2100 2111 0 0 0
9110 9119 5114 0 0
1113 5676 4875 2186 1153
6220 6225 1293 0 0
1115 3234 2163 0 0
7110 7873 0 0 0
1100 2679 8153 2114 1145"
)
it <- mutate_all(it, as.character)
matchit <- function(x){
tmp <- x[-1]
mypattern = paste0("^",stringr::str_sub(x[[1]],1,2),".*$")
hit <- tmp[which(grepl(mypattern, tmp))]
return(ifelse(length(hit), hit[[1]], x[[1]]))
}
it$update <- apply(it, 1, matchit)
it
#> NAICS_2017 NAICS 2012_1 NAICS 2012_2 NAICS 2012_3 NAICS 2012_4 update
#> 1 2100 2111 0 0 0 2111
#> 2 9110 9119 5114 0 0 9119
#> 3 1113 5676 4875 2186 1153 1153
#> 4 6220 6225 1293 0 0 6225
#> 5 1115 3234 2163 0 0 1115
#> 6 7110 7873 0 0 0 7110
#> 7 1100 2679 8153 2114 1145 1145
Explanation:
The function returns the first match of the pattern in a row(except the 1st element), if not, it returns the 1st element,
Then we can update the data.frame by applying it to each row.