Generate Mermaid ERD from DuckDB Connection

Generates a Mermaid entity relationship diagram from a DuckDB connection by querying information_schema.columns. Unlike dm::dm_draw(), this handles GEOMETRY columns without errors, so spatial tables like site, grid, casts, and segment are included in the diagram.

cc_erd(
  con,
  tables = NULL,
  exclude = NULL,
  rels_path = NULL,
  rels = NULL,
  colors = NULL,
  layout = "elk",
  view_type = "all"
)

Arguments

con: DBI connection to a DuckDB database
tables: Character vector of table names to include. If NULL (default), all tables in the connection are included.
exclude: Character vector of table names to exclude (default: NULL). Applied after tables.
rels_path: Path to a relationships.json file for primary key and foreign key definitions. If NULL (default), the diagram shows table structures without relationship lines. Ignored when rels is provided.
rels: A list with primary_keys (named list: table → column) and foreign_keys (list of lists with table, column, ref_table, ref_column). Alternative to rels_path for passing relationships inline. Takes precedence over rels_path.
colors: Named list mapping color names or hex codes to character vectors of table names. Emitted as Mermaid classDef/class directives with a darker auto-generated stroke. Example: list(lightblue = c("cruise", "ship")).
layout: Layout engine: "elk" (default) or "dagre".
view_type: What columns to show: "all" (default), "keys_only" (PK/FK columns only), or "title_only" (table names, no columns).

Value

An object of class "cc_erd" (inherits "character"). Printing outputs the Mermaid code; in Quarto/RMarkdown it auto-renders as a Mermaid diagram (respects mermaid-format in _quarto.yml).

Examples

if (FALSE) { # \dontrun{
con <- cc_get_db()

# basic ERD with all tables (including geometry tables)
cc_erd(con)

# with relationships from frozen release
cc_erd(con, rels_path = "data/releases/v2026.03.25/relationships.json")

# color-coded groups matching release_database.qmd palette
cc_erd(con,
  rels_path = "relationships.json",
  colors = list(
    lightblue   = c("cruise", "ship", "site", "tow", "net"),
    lightyellow = c("ichthyo", "species", "lookup", "taxon", "taxa_rank"),
    lightgreen  = c("grid", "segment"),
    pink        = c("casts", "bottle", "bottle_measurement",
                     "cast_condition", "measurement_type"),
    lavender    = c("ctd_cast", "ctd_measurement", "ctd_summary"),
    lightsalmon = c("dic_sample", "dic_measurement",
                     "dic_measurement_summary"),
    white       = c("dataset")))

# inline relationships (alternative to rels_path)
cc_erd(con, rels = list(
  primary_keys = list(cruise = "cruise_key", ship = "ship_key"),
  foreign_keys = list(
    list(table = "cruise", column = "ship_key",
         ref_table = "ship", ref_column = "ship_key"))))

# compact: show only key columns
cc_erd(con, rels_path = "relationships.json", view_type = "keys_only")
} # }