Construct a field-level data-normalization concordance

Creates field-level rules for interpreting blank source values. Data normalization is distinct from panel imputation: it interprets source encoding only and never creates missing organization-year rows.

Called with no arguments, concordance() returns the package's built-in concordance for IRS 990 efile variables, derived from the IRS Efile Master Concordance File (see field_concordance). Supply field and blank_meaning to build a custom concordance instead.

Usage

concordance(
  field = NULL,
  blank_meaning = NULL,
  forms = "*",
  table = NA_character_,
  notes = NA_character_
)

Arguments

field

Character vector of source field names. NULL (the default) returns the built-in 990 concordance.

blank_meaning

Character vector containing "implicit_zero", "implicit_false", or "literal_missing".

forms

A character vector or list. Each rule supplies the applicable return types, such as c("990", "990EZ"). Character entries may use | as a separator. Use "*" for every form.

table

Optional source-table name for auditing.

notes

Optional explanatory notes.

Value

A concordance data frame with a list-column named forms.

See also

field_concordance for the underlying data, fields_in_scope() to select fields by form scope, and normalize() to apply the rules.