TopolearnSubsetting is how you filter and reshape every kind of data in R.
R indexes from 1. The [ operator accepts four kinds of index:
- Positive integers select those positions:
x[c(1, 3)]. Out-of-range positions giveNA. - Negative integers drop positions:
x[-1]is everything but the first. You can't mix positive and negative. - Logical vectors keep elements where the index is
TRUE:x[x > 10]. This is how you filter. - Character vectors select by name:
x[c("a", "b")]for a named vector.
Name elements with c(a = 1, b = 2) or names(x) <- c("a", "b").
Subsetting works on the left of an assignment too: x[x < 0] <- 0 replaces every negative value in one step, no loop needed. Assigning beyond the end extends the vector, filling gaps with NA.
which(cond) turns a logical vector into the positions that are TRUE, and which.max(x) finds the position of the maximum. %in% tests membership: x[x %in% c(2, 4)].
x[0] is an empty vector, not the first element, which trips up people coming from zero-based languages. Missing values in a logical index produce NA elements, so filter with x[!is.na(x) & x > 10] or x[which(x > 10)] when data may contain NA.
Example
x <- c(a = 5, b = -2, c = 9, d = -7)
print(x[c(1, 3)]) #> a c
#> 5 9
print(x[-1]) #> b c d / -2 9 -7
print(x[x > 0]) #> a c / 5 9
print(x[c("d", "a")]) #> d a / -7 5
print(which(x < 0)) #> b d / 2 4
x[x < 0] <- 0 # replace negatives in place
print(unname(x)) #> [1] 5 0 9 0
y <- c(3, NA, 12)
print(y[y > 5]) #> [1] NA 12
print(y[which(y > 5)]) #> [1] 12
print(length(x[0])) #> [1] 0Remember
- Indexes start at 1; negative indexes drop elements
- A logical index filters:
x[x > 0] - Assign into a subset to modify many elements at once
- Use
which()or!is.na()when NAs may be present
Watch out for
- Using
x[0]for the first element, which returns an empty vector.