Skip to contents

Recode Values by Ordered Conditions

Source: R/caseWhen.R
caseWhen.Rd

Vectorized conditional selection: the first matching rule wins.

Usage

caseWhen(..., default = NULL, na = c("propagate", "skip"))

Arguments

...

two-sided formulas of the form condition ~ value. Conditions must be plain logical vectors. Values must be plain logical, integer, double, complex, or character vectors. Classed objects (including factors and dates), arrays, and lists are not supported.

default

value for elements without a match. NULL (the default) supplies a missing value of the common result type.

na

handling of missing conditions: "propagate" (default) resolves an element to a missing value when its first unresolved condition is NA; "skip" treats missing conditions as FALSE.

Value

An unnamed atomic vector of the common length and type.

Details

Unlike dplyr::case_when(), which treats missing conditions as FALSE, this function defaults to propagating NA, as in nested ifelse() calls. Use na = "skip" to skip missing conditions instead.

All conditions and values are evaluated once in their formula environments, including values of unused rules. This is ordered selection, not lazy or elementwise short-circuit evaluation. Errors and warnings in unused expressions therefore still occur.

All inputs (including default) must have length one or the common output length. Non-scalar lengths must agree; zero-length vectors are supported but cannot be combined with lengths greater than one. All-scalar inputs produce length one. Names on inputs are discarded.

Numeric types are promoted in the order logical, integer, double, complex. Character values cannot be mixed with numeric or non-missing logical values. All-missing logical vectors are treated as unspecified missing values and may accompany any supported type. Every rule participates in length and type validation, even if it never matches.

Factors, dates, and other classed vectors are deliberately unsupported in this version; support may be added later with explicit class and factor-level compatibility rules.

See also

ifelse(), dplyr::case_when()

Other data.recode: asBinary(), combLevels(), dummy(), mReplace(), nf(), recodeX(), revCode(), stringsAsFactors()

Examples

x <- c(-1, 0, 2, NA_real_)
caseWhen(x < 0 ~ "negative", x == 0 ~ "zero", default = "positive")
#> [1] "negative" "zero"     "positive" NA        
caseWhen(x < 0 ~ "negative", default = "other", na = "skip")
#> [1] "negative" "other"    "other"    "other"   
caseWhen(c(TRUE, FALSE, NA) ~ 1L, default = 0L)
#> [1]  1  0 NA