Skip to contents

Join values with a separator, skipping the missing ones

Source: R/catx.R
catx.Rd

Concatenates its arguments element by element, separated by sep, leaving out every value that is missing (NA) or empty (""). This is the SAS CATX rule, and it is what listing_col() applies to its vars: a record whose middle value is missing prints "COMPLETED/HYPERGLYCAEMIA", never "COMPLETED//HYPERGLYCAEMIA".

Usage

catx(sep, ...)

Arguments

sep

Separator, a single string. It is inserted only between values that survive, so it never appears at the start or the end.

...

The values to join: character vectors, or anything as.character() accepts. A factor joins by its labels, not its level codes.

Value

A character vector as long as the longest argument. With no ... arguments, character(0).

Details

Arguments are vectorised together. A length-1 argument is recycled to the longest one; any other length mismatch is an error, because in a clinical listing it means two columns that were meant to line up do not.

Leading and trailing blanks are stripped from each value before it is joined – again as SAS CATX does. Source data read from a SAS-shaped export is routinely padded, and padding that survived into a joined cell would push the text over the column width for no visible reason. (This is the one place the function differs from ydisctools::catx(), which keeps the padding.)

See also

listing_col(), which joins a listing column's vars this way; build_listing(), which applies it.

Examples

catx("/", "COMPLETED", "UPPER", "HYPERGLYCAEMIA")
#> [1] "COMPLETED/UPPER/HYPERGLYCAEMIA"

# A missing value is skipped, not printed as a doubled separator.
catx("/", "COMPLETED", NA, "HYPERGLYCAEMIA")
#> [1] "COMPLETED/HYPERGLYCAEMIA"

# Vectorised, with length-1 arguments recycled.
adsl <- data.frame(
  AEDECOD = c("HYPERGLYCAEMIA", "SKIN ULCER"),
  AELOC   = c("UPPER", NA),
  stringsAsFactors = FALSE
)
catx("/", adsl$AEDECOD, adsl$AELOC)
#> [1] "HYPERGLYCAEMIA/UPPER" "SKIN ULCER"          
catx(" ", "Cohort", c("A", "B"))
#> [1] "Cohort A" "Cohort B"

# Any separator, including none at all.
catx(", ", "Grade 3", "Prior radiation")
#> [1] "Grade 3, Prior radiation"
catx("", "AE", "0001")
#> [1] "AE0001"