Add absent columns to a tibble and safely cast present columns to the types declared by a zero-row prototype. Missing columns receive typed missing values unless a scalar default is supplied. Defaults never replace missing values inside a column that is already present.
Prototype columns are returned first in prototype order. With
extra = "keep", other input columns follow in their original relative
order. With extra = "drop", the output has exactly the prototype schema.
Usage
ensure_columns(data, prototype, defaults = NULL, extra = c("keep", "drop"))Arguments
- data
Input tibble.
- prototype
Zero-row tibble defining required column names, types, and order.
- defaults
Optional named list of size-one defaults for absent columns. Every name must occur in
prototype, and each value must be safely convertible to the corresponding prototype type.- extra
Whether columns absent from
prototypeare"keep"or"drop".
Value
An ungrouped tibble with the same rows as data, completed and
ordered from prototype. Grouping metadata is not preserved.
Examples
prototype <- tibble::tibble(
year = integer(),
value = double(),
source = character()
)
data <- tibble::tibble(value = 2, year = 2020L, note = "observed")
ensure_columns(
data,
prototype,
defaults = list(source = "unknown")
)
#> # A tibble: 1 × 4
#> year value source note
#> <int> <dbl> <chr> <chr>
#> 1 2020 2 unknown observed