search and replace strings, not matching pattern, and NA in R

Question

I have character vector (column) in a data.frame that contains three correct values comprised of "Positive", "Negative" and NA, along a small number of incorrect values. I want to replace the incorrect values, i.e. everything except "Positive", "Negative" and "NA" with "Positive" I can grep for NOT "Positive" or "Negative", using grep and the invert=TRUE argument, and get everything except instances of "Negative" and "Positive" but I've not figured out how to additionally, not replace any instances of NA. I've also looked at the sub command, but it does not have an invert option. Any suggestions?

Please read the info about how to give a [minimal reproducible example](http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example/5963610). — Jaap, Oct 14 '15 at 15:01

score 3 · Accepted Answer · edited Oct 14 '15 at 15:37

3

This should work:

df[!is.na(df$column) & !df$column %in% 
        c("Positive","Negative"), "column"] <- "Positive"

edited Oct 14 '15 at 15:37

akrun

874,273
37
540
662

answered Oct 14 '15 at 15:00

Jaap

81,064
34
182
193

search and replace strings, not matching pattern, and NA in R

1 Answers1