CODEWITHSUNDEEP
rconditional-statementsmissing-datana
Matt Bannert
Matt Bannert

Reputation: 28264

How to handle conditional data.frame subscripts that contain NAs?

Is there a way to have conditional subscripts if the conditional vector contains NAs? Assume i have a data.frame like this

dframe <- data.frame(a=c(1,32,4,5,8),b=c(1,2,3,4,5),d=c(NA,5,5,10,9))
dframe[dframe$d > 9,"a"] <- NA

If it was not for the NA in dframe$d this would be straight forward. I have seen %in% syntax like here to get around NAs, but do not know how to manage it for conditions. I can see that this is somewhat of a general problem, since I am not quite sure whether I want to obtain an NA for the missing value in the condition or something else. But I am also interested to learn how people handle this situation.

In my specific situation in would simply be helpful when NA was treated like FALSE in the condition.

Upvotes: 4

Views: 3737

Answers (2)

kith
kith

Reputation: 5566

You can just restrict your indexing to the non NA values

dframe[ dframe$d > 9 & !is.na(dframe$d), "a"] <- NA

Edit: Here is a more compact alternative from the R Inferno:

dframe[ which(dframe$d > 9), "a"] <- NA

Upvotes: 6

Arun
Arun

Reputation: 118799

You were close. You had the condition and %in%. Just not the link. It is straightforward actually. Just use %in% with TRUE it'll give back all other values, where as NA will be replaced with FALSE

 dframe[(dframe$d > 9) %in% TRUE, "a"] <- NA

Upvotes: 5

Related Questions