Compare one variable to other variables by group in R - r

Question

I have the following data frame:
data.frame(id = c("a", "a", "a", "d", "d"),
value = c(5, 46, 12, 14, 32),
low = c(46, 8, NA, 0, 34),
high = c(56, 20, NA, 12, 60))
id value low high
1 a 5 46 56
2 a 46 8 20
3 a 12 NA NA
4 d 14 0 12
5 d 32 34 60
I need to set a new variable to TRUE if value is out of every intervals defined by low and high for each line with the same id.
My desired dataframe would be:
id value low high result
1 a 5 45 56 TRUE # 5 not in 45-56, 8-20
2 a 46 8 20 FALSE # 46 in 45-56
3 a 12 NA NA FALSE # 12 in 8-20
4 d 14 0 12 TRUE # 14 not in 0-12, 34-60
5 d 32 34 60 TRUE # 32 not in 0-12, 34-60
How can I do it in base R? I work in a restrictive environment where I only have access to base R.

Answer 1

I figured out an ugly and not optimized solution but it works ! Here is the code :
df <- data.frame(id = c("a", "a", "a", "d", "d"),
value = c(5, 46, 12, 14, 32),
low = c(46, 8, NA, 0, 34),
high = c(56, 20, NA, 12, 60))
list.inter <- list()
for(i in 1:nrow(df)){
if(is.na(df$low[i]) | is.na(df$low[i])) {
list.inter[[i]] <- NA
}else{
list.inter[[i]] <- seq(from = df$low[i], to = df$high[i])
}
}
result <- c()
for(i in 1:nrow(df)){
result[i] <- ! df$value[i] %in% unlist(list.inter[which(df$id[i]==df$id)])
}
df$result <- result
I hope it helps and I am curious to see some optimized code from other users!

Answer 2

Without apply, sapply and map function:
isInDataframe <- function(data = data, value = "value", from = "low", to = "high", id = "id"){
result <- c()
for (i in 1:length(data[,1])) {
deeta <- data[data[id] == as.character(data[id][i,1]),]
subresult <- c()
for (j in 1:nrow(deeta)) {
subresult[j] <- (data[value][i,1] >= deeta[from][j,1] & data[value][i,1] <= deeta[to][j,1])
}
result[i] <- !any(subresult,na.rm = T)
}
data$result <- result
return(data)
}
isInDataframe(data = data, value = "value", from = "low", to = "high", id = "id")
id value low high result
1 a 5 46 56 TRUE
2 a 46 8 20 FALSE
3 a 12 NA NA FALSE
4 d 14 0 12 TRUE
5 d 32 34 60 TRUE

Answer 3

I finally choose to separate id and value in a data frame and id, low and high in another data frame for this analysis.
However, here is a solution highly inspired from the solutions suggested for this new approach:
df <- data.frame(id = c("a", "a", "a", "d", "d"),
value = c(5, 46, 12, 14, 32),
low = c(46, 8, NA, 0, 34),
high = c(56, 20, NA, 12, 60))
temp <- merge(x = df[c("id",
"value")],
y = df[c("id",
"low",
"high")])
temp$result <- temp$value < temp$low | temp$value > temp$high
merge(x = df,
y = aggregate(formula = result ~ id + value,
data = temp,
FUN = all))
id value low high result
1 a 12 NA NA FALSE
2 a 46 8 20 FALSE
3 a 5 46 56 TRUE
4 d 14 0 12 TRUE
5 d 32 34 60 TRUE