Pebellslearn by building
0%
RoadmapsMatrixCoursesSandboxSign in
Data Types  ›  Lesson

Logical — TRUE, FALSE and NA

Logical 12 minLogical values are TRUE and FALSE, which count as 1 and 0 in arithmetic, plus NA for an unknown value that spreads through calculations unless you remove it or test for it with is.na().

Logical — TRUE, FALSE and NA

Plot F's harvest was never weighed. Suddenly the average weight comes back as NA, and a filter that worked yesterday behaves strangely. Logical values drive every filter and decision in R, and their third value, NA, is where most surprises come from.

The rule

TRUE counts as 1, FALSE as 0, and NA means unknown.

Logical values come from comparisons, or you type them directly:

watered <- c(TRUE, FALSE, TRUE, TRUE, FALSE)

sum(watered)     # 3, the number of watered plots
mean(watered)    # 0.6, the share of plots watered
any(watered)     # TRUE, at least one is TRUE
all(watered)     # FALSE, not every one is TRUE
which(watered)   # 1 3 4, the positions of the TRUEs

Because TRUE counts as 1 and FALSE as 0, sum() of a logical vector is a count and mean() is a proportion. This is one of R's most useful habits.

Always type TRUE and FALSE in full. T and F usually work, but they are ordinary variables that anyone can overwrite, for example with T <- 0, while TRUE and FALSE are reserved words.

NA is "not available": a value exists but nobody knows it. It is logical by default (typeof(NA) is "logical") and it can sit in any vector:

kg <- c(12.5, 8, 15, 3.5, 6, NA)   # plot F was not weighed

mean(kg)                 # NA
mean(kg, na.rm = TRUE)   # 9
kg > 7                   # TRUE TRUE TRUE FALSE FALSE NA
kg == NA                 # NA NA NA NA NA NA
is.na(kg)                # FALSE FALSE FALSE FALSE FALSE TRUE

NA spreads: an unknown value makes any result that depends on it unknown. Many summary functions accept na.rm = TRUE to drop NAs first. To find NAs, always use is.na(), because x == NA asks "is x equal to an unknown value?", and the only honest answer is NA.

Do not confuse NA with NULL. NA is a placeholder for one missing value (length(NA) is 1); NULL means nothing at all (length(NULL) is 0).

The shortcut

For any expression involving NA, ask: "Could the unknown value change the answer?" If yes, the result is NA. If no, R gives the definite answer.

  • NA & FALSE is FALSE: whatever NA is, "and FALSE" is FALSE.
  • NA & TRUE is NA: the answer depends on the unknown.
  • NA + 1 is NA: an unknown number plus one is still unknown.
Counting heavy plots with a missing weightStep 1 of 4

  1. Compare

    kg > 7 gives TRUE TRUE TRUE FALSE FALSE NA. Plot F's comparison is unknown.

Drills

Answer fast — trust the rule.

quick drill

With watered <- c(TRUE, FALSE, TRUE, TRUE, FALSE), what is sum(watered)?

quick drill

What is mean(watered)?

quick drill

What does NA == NA return?

quick drill

What does NA & FALSE return?

quick drill

Which is the correct way to find the missing values in kg?

Remember

  • TRUE counts as 1, FALSE as 0, and NA means unknown.
  • Shortcut: could the unknown change the answer? If so, the result is NA.
  • Trap: testing with == NA, forgetting na.rm = TRUE, and trusting T and F.

Practice — level up

0/1 solved

Write your answer, then compare it with a rubric and a model answer.

ExplainThe missing weight

Plot F's harvest was never weighed, so the vector is kg <- c(12.5, 8, 15, 3.5, 6, NA). Explain what mean(kg), mean(kg, na.rm = TRUE), kg == NA and is.na(kg) each return, and why.