Check data harmonization

Functions to check that data tables (trait tables or allometry tables) have the appropriate format and data for model parameter filling.

Usage

check_harmonized_trait(x, verbose = TRUE)

check_harmonized_trait_dir(
  harmonized_trait_path,
  update_version = FALSE,
  verbose = TRUE
)

check_harmonized_allometry(x, verbose = TRUE)

check_harmonized_allometry_dir(
  harmonized_allometry_path,
  update_version = FALSE,
  verbose = TRUE
)

Arguments

x

A data frame with harmonized trait data or harmonized allometry data.

verbose

A logical flag indicating extra console output (information alerts and check success). If FALSE, only errors are reported.

harmonized_trait_path

The path to harmonized trait data files (.rds or .csv format).

update_version

If TRUE and the package is acceptable, it updates the package version and stores the harmonized file.

harmonized_allometry_path

The path to harmonized allometry data files (.rds or .csv format).

Value

• Functions check_harmonized_trait() and check_harmonized_allometry() return an (invisible) logical indicating whether the data frame is acceptable or not.

• Functions check_harmonized_trait_dir() and check_harmonized_allometry_dir() return an (invisible) logical vector indicating those files whose data is acceptable or not.

Details

Function check_harmonized_trait() checks that the input data frame conforms to the following requirements:

1. Has columns called originalName, acceptedName, acceptedNameAuthorship, family, genus, specificEpithet, and taxonRank, as returned by function harmonize_taxonomy_WFO.

2. Trait data can be in wide-format or long-format. When supplied in wide format, the names of trait columns are either a valid trait names according to the notation in HarmonizedTraitDefinition and the type of column needs to match the corresponding type. In this format trait units cannot be checked and the same reference applies to all traits. When supplied in long format, the following columns are required:

   • Trait - A character describing the the name of the trait, with values according to the notation in HarmonizedTraitDefinition.

   • Value - A character or numeric containing the trait value. For numeric traits, values should be coercible to numbers via as.numeric.

   • Units - A character describing the units of the trait, with values according to units in HarmonizedTraitDefinition.

3. The names of the remaining columns are:

   • Reference - A string describing the harmonized trait data source, i.e. normally a reference.

   • DOI - A string describing the DOI of the harmonized trait data source.

   • OriginalReference - A string describing the original source of trait data (if Reference is a compilation).

   • OriginalDOI - A string describing the DOI of the original source of trait data (if Reference is a compilation).

   • Priority - A numeric value to prioritize some data values (sources) over others (1 means highest priority).

   • checkVersion - A string of the package version used for harmonization checking.

   Columns OriginalReference and OriginalDOI are optional, and the remaining columns are recommended.

4. Trait value columns contain non-missing data.

5. Numeric trait columns do not contain values beyond expected ranges (when defined).

6. Trait columns do not contain values different than accepted values (when defined, typically for categorical traits). Lower-upper cases are ignored.

Function check_harmonized_allometry() checks that the input data frame conforms to the following requirements:

1. Has columns called originalName, acceptedName, acceptedNameAuthorship, family, genus, specificEpithet, and taxonRank, as returned by function harmonize_taxonomy_WFO.

2. Has columns called Response and Equation. Columns Priority, Reference and DOI are recommended.

3. Column Response identifies the response variable.

4. Column ResponseDescription contains a longer description of the response variable.

5. Columns Predictor1, Predictor2, ... identify predictor variables.

6. Columns PredictorDescription1, PredictorDescription2, ... contain longer descriptions of predictor variables.

7. Allometry coefficient columns should be named a, b, c, ..., f and contain numeric values.

Functions check_harmonized_trait_dir() and check_harmonized_allometry_dir() allow checking the appropriateness of multiple harmonized files at once.

See also

harmonize_taxonomy_WFO, HarmonizedTraitDefinition