Package 'hrmn'

Title: Harmonize Datasets
Description: A common early step during data analysis is "data harmonization" -- converting disparate datasets into a unified, consistent format, with consistent column names, classes, and values. The goal of 'hrmn' is to make this process as easy as it can be.
Authors: Jon Harmon [aut, cre] (ORCID: <https://orcid.org/0000-0003-4781-4346>)
Maintainer: Jon Harmon <[email protected]>
License: MIT + file LICENSE
Version: 0.0.0.9005
Built: 2026-05-29 17:06:36 UTC
Source: https://github.com/wranglezone/hrmn

Help Index

Harmonize a data frame

Description

Harmonize a data frame

Usage

harmonize_df(
  .data,
  ...,
  .spec = NULL,
  .unspecified_columns = c("error", "drop", "keep")
)

Arguments

.data

(data.frame) A data frame to harmonize.
...

These dots are for future extensions and must be empty.
.spec

(hrmn_spec_df) A data frame harmonization specification.
.unspecified_columns

("error", "drop", or "keep") How to handle columns in .data that are not present in .spec.

Value

The input .data harmonized to a tibble::tibble().

See Also

Other harmonization functions: harmonize_fct()

Examples

df <- data.frame(
  size = c("Small", "Medium", "S", "M", "Large", "Lrg", "Sm"),
  id = 1:7
)

# This spec will coerce values to NA if they are not "Small", "Medium",
# or "Large".
spec <- specify_df(
  size = specify_fct(levels = c("Small", "Medium", "Large"))
)

# We can provide harmonization rules to the data before the spec is applied.
# Here, we harmonize the input factor to convert "S", "M", "Sm", and "Lrg" to
# valid values.
harmonize_df(
  df,
  size = harmonize_fct(
    size,
    .lookup = c("S" = "Small", "M" = "Medium", "Sm" = "Small", "Lrg" = "Large")
  ),
  .spec = spec,
  .unspecified_columns = "keep"
)

Harmonize a factor

Description

Harmonize a factor

Usage

harmonize_fct(.data, ..., .spec = NULL, .lookup = NULL)

Arguments

.data

(character or coercible to character) A vector to harmonize to the specified factor.
...

These dots are for future extensions and must be empty.
.spec

(hrmn_spec_fct) A harmonization specification from specify_fct().
.lookup

(named character) A vector of replacement values. The names are the values in .data and the values are the target values.

Value

A harmonized factor().

See Also

Other harmonization functions: harmonize_df()

Examples

# Without a spec, harmonize_fct() acts like [base::factor()].
harmonize_fct(c("a", "b", "c"))

# Basic harmonization, dropping levels not in the spec
spec <- specify_fct(levels = c("a", "b"))
harmonize_fct(c("a", "b", "c"), .spec = spec)

# Using a lookup table to recode values
spec2 <- specify_fct(levels = c("fruit", "citrus"))
lookup <- c(apple = "fruit", banana = "fruit", orange = "citrus")
harmonize_fct(
  c("apple", "banana", "orange"),
  .spec = spec2,
  .lookup = lookup
)

Data frame specification

Description

Create an object that specifies the desired format for a data frame. This specification object does not contain any data itself, only the rules for harmonization.

Usage

specify_df(...)

Arguments

...

(hrmn_spec) Column specifications, given as named arguments.

Value

A hrmn_spec_df object that acts as a specification.

See Also

Other specification functions: specify_fct()

Examples

specify_df(
  response = specify_fct(levels = c("Yes", "No", "Maybe")),
  outcome = specify_fct(levels = c("Positive", "Negative"))
)

Factor specification

Description

Create an object that specifies the desired levels for a factor variable. This specification object does not contain any data itself, only the rules for harmonization.

Usage

specify_fct(levels = character())

Arguments

levels

(character) The allowed values of the factor.

Value

A hrmn_spec_fct object that acts as a specification.

See Also

Other specification functions: specify_df()

Examples

specify_fct(levels = c("a", "b", "c"))