Append event rows into the core sample dimension — append

select_sql is bound positionally, so it must yield either the 15 columns of the base contract — sample_key, sample_type, parent_sample_key, root_sample_key, dataset_key, grid_key, site_key, cruise_key, order_occ, latitude, longitude, datetime, depth_min_m, depth_max_m, tow_type — or those 15 plus a trailing 16th, data_stage. geom is minted here as ST_Point(longitude, latitude). tow_type is the net gear code (ichthyo tow/net grains: C1/CB/CV/PV oblique/vertical, MT manta), NULL for gears/datasets without one. Call it once per event level — a multi-level dataset (ichthyo site->tow->net, bottle cast->bottle) appends one arm per level, and sample_arm_self() writes the single-level case for you.

Usage

append_sample(con, select_sql, sample_tbl = "sample")

Arguments

con: a DuckDB connection (open via get_duckdb_con())
select_sql: a SELECT producing the canonical obs columns by name
sample_tbl: target table (default "sample")

Value

(invisibly) the total row count of sample_tbl after the append

Details

data_stage is optional and trailing on purpose: it records the source's own processing state for the event (final vs preliminary for CTD casts, per question calcofi_ctd-cast_14), which most datasets do not distinguish. Making it positional column 16 rather than inserting it into the contract lets a dataset opt in when it has a meaningful stage without touching the other arms — a 15-column arm gets NULL and keeps working unchanged.