Skip to contents

Guess a column's type from its values and its name

Source: R/02-02-classify-by-name.R
guess_column.Rd

Runs both detection modalities and reconciles them VALUES-FIRST: if a value detector clears threshold, its (format-anchored) answer is returned; only when no value detector is confident does the name classifier fill in the class. This is what lets the metadata-gated classes – a double column named sampling_weight, a 0/1 column named eligible – be typed at all, while never letting a misleading name override a real value signature.

Usage

guess_column(x, name = NULL, threshold = 0.8, ...)

Arguments

x

A vector (coerced to character) of column values.

name

Optional variable name (header string).

threshold

Minimum value-detector pass rate to trust the value guess.

...

Passed to guess_data_type().

Value

A list with source ("value", "name", or "none"), guess, ontology (the standard 4-vector), confidence, and the full value and name sub-results for inspection.

See also

guess_data_type(), classify_by_name()

Examples

guess_column(c("06037", "36061"), name = "county")$source    # "value"
#> Warning: NAs introduced by coercion to integer range
#> [1] "value"
guess_column(c("3.4", "5.1", "7.8"), name = "sampling_weight")$guess  # weight
#> Warning: NAs introduced by coercion to integer range
#> [1] "weight"