Editable flat columns of the plots or individuals table — .upd_direct

Deliberately an allow-list, not everything the schema happens to hold. `data_individuals` and `data_liste_plots` both carry deprecated columns (`dbh`, `code_individu`, `sous_plot_name` and others) that nothing writes any more; offering them for editing would invite corrections that change no behaviour, and in `dbh`'s case would silently disagree with the `stem_diameter` measurements in `data_traits_measures`.

Usage

.upd_direct_fields(entity = c("plot", "individual"), con)

Arguments

entity: `"plot"` or `"individual"`.
con: A DBI connection.

Value

A tibble with one row per editable column: `field`, `pg_type`, `kind` (`"text"`, `"numeric"`, `"integer"`, `"boolean"`, `"lookup"`, `"taxon"`), and for lookups `lookup_table`, `lookup_key`, `lookup_value`. Carries a `hidden` attribute listing the omitted schema columns.

Details

`get_table_columns()` is the package's maintained answer to "which columns of this table does a user actually set", already used by the import wizard and by `update_records()`. This intersects it with the live schema, so a name it lists that is not a real column (`plot_name` on individuals, which is reached through `id_table_liste_plots_n`) drops out rather than producing a form field that cannot be written.

The friendly lookup names `method` and `country` are swapped for the id columns that store them, and flagged `"lookup"` so the UI offers a dropdown. `idtax_n` is flagged `"taxon"` so it gets a taxonomic search instead of a raw id box.

Columns the schema has but the allow-list omits are recorded in the `hidden` attribute, so the UI can say what it is not showing.