Build Efficient R Filter - r

Question

I have this dataframe called data. In the data frame I have a few columns, for simplicity I will explain the columns with a weather analogy, it is like "weather_st_louis", "weather_boston", "weather_ny"... I want to build a column "weather" and it should be done like this, "if weather in st louis exists, use this column, else if weather in boston exists, use this column, else if weather in ny exists, use this column, else NONE". I'm going to be using this logic many times, with many columns, so need a way to make this more efficient. What is the R way to do this.
Also, side question, is what I'm trying to build here called a "filter"?
if(exists("data['w_stlouis']")) {
data['w'] <- data['w_stlouis']
} else if(exists("data['w_boston]")){
data['w'] <- data['w_boston']
} else if(exists("data['w_ny']")){
data['w'] <- data['w_ny']
} else {data['w'] <- NA}

Answer 1

Try something like that :
example <- matrix(NA,ncol=5,nrow=5)
colnames(example) <- c("weather_1","weather_2","weather_3","weather_4","weather_5")
example[5,3] <- 1
example[3,2] <- 1
example[1,2] <- 1
example[4,4] <- 1
example[5,2] <- 1
w <- apply(example,1,function(x){
o <- which(!is.na(x))[1]
if (is.na(o)) r <- "NONE"
else r <- colnames(example)[o]
r
})
w
When you have repeating tasks to do, try to use apply/tapply/sapply functions

Answer 2

Here is another posibility. I'm not sure if this is what you need, but perhaps it gives you another way to handle it.
df <- data.frame(matrix(rnorm(100, 100, 20),ncol=5,nrow=5))
colnames(df) <- c("weather_1","weather_2","weather_3","weather_4","weather_5")
library(reshape2)
df <- melt(df)
df[1:10,2] <- NA
str(df)
weather_levels <- levels(df$variable)
df$case <- ifelse(is.na(df$value), 0, 1)
These two output the same result
subset(df, df$case == 1)
na.omit(df)