Skip to contents

Remove rows from one or more population tables

Source: R/remove_rows.R
remove_rows.Rd

Removes rows from the database based on a filtered tidybreed_table. Chain after get_table() and dplyr::filter() to target specific rows. Returns the population invisibly — consistent with all other action functions.

Usage

remove_rows(
  tbl,
  tables = NULL,
  confirm_all = FALSE,
  dry_run = FALSE,
  verbose = TRUE
)

Arguments

tbl

A tidybreed_table from get_table(), with a dplyr::filter() applied.

tables

NULL (default) to delete from the current table only; "all" to delete from every ind_* table (including ind_haplotype and ind_genotype) that exists in the population; or a character vector of specific table names (each must have an id_ind column). When not NULL, the source table must also have an id_ind column.

confirm_all

Logical. Set to TRUE to allow deletion of all rows when no filter has been applied. This is a deliberate safeguard — you must explicitly opt in to wiping an entire table. Default FALSE.

dry_run

Logical. If TRUE, report what would be deleted without modifying the database. Default FALSE.

verbose

Logical. If TRUE, print a message after each deletion reporting the row count removed. Default TRUE.

Value

The tidybreed_pop object (invisibly).

Details

A filter is required by default. Calling remove_rows() without a prior dplyr::filter() will stop with an error explaining how to opt in to full-table deletion via confirm_all = TRUE.

Single-table mode (tables = NULL): uses the composite row key from the internal TABLE_ROW_KEYS registry to delete exactly the rows matched by the filter — no more, no less. For example, filtering ind_tbv by trait_name == "ADG" deletes only the ADG rows, not all TBV rows for those animals. Key columns are matched with IS NOT DISTINCT FROM, so rows whose key is NULL — the default chr_inheritance / chr_recombination rows seeded by define_genome(), or the shared (line_name IS NULL) founder_haplotypes pool — delete correctly.

Every system table is registered for single-table deletion except _schema_meta, which is package-managed; remove_rows() refuses it with a pointer to define_schema_description() rather than a generic error.

Cross-table mode (tables != NULL): extracts unique id_ind values from the filtered table and issues a DELETE ... WHERE id_ind IN (...) for each target table via a temp-table JOIN. tables = "all" targets every ind_* table (including ind_haplotype, ind_genotype, and ind_true_index) that currently exists in the population (including ind_meta).

A temporary DuckDB table is used for both modes to avoid SQL injection risk and to handle large ID sets efficiently.

Examples

if (FALSE) { # \dontrun{
# Delete specific phenotype records for culled animals
pop |>
  get_table("ind_phenotype") |>
  dplyr::filter(id_ind %in% culled_ids, phenotype_name == "litter_size") |>
  remove_rows()

# Remove all data for culled animals across every individual table
pop |>
  get_table("ind_meta") |>
  dplyr::filter(id_ind %in% culled_ids) |>
  remove_rows(tables = "all")

# Delete from an explicit subset of tables only
pop |>
  get_table("ind_meta") |>
  dplyr::filter(id_ind %in% culled_ids) |>
  remove_rows(tables = c("ind_phenotype", "ind_tbv"))

# Preview before deleting
pop |>
  get_table("ind_meta") |>
  dplyr::filter(id_ind %in% culled_ids) |>
  remove_rows(tables = "all", dry_run = TRUE)

# Wipe an entire table (requires confirm_all = TRUE)
pop |>
  get_table("ind_phenotype") |>
  remove_rows(confirm_all = TRUE)
} # }