One physical CSV file in a dataset version.
Ingestion emits one row per CSV on disk. SSC repeats a measure across role
folders and so yields one row per (table, role); SPARK yields one per table.
-
dataset
Dataset id.
- Type:
str
-
version
Dataset version.
- Type:
str
-
table_name
Logical table name (the CSV stem).
- Type:
str
-
display_title
Human-facing table title from the dictionary, when available.
- Type:
str
-
role
Family role for SSC (proband, mother, and so on); empty for SPARK.
- Type:
str
-
file_path
POSIX path to the CSV, relative to the repository root.
- Type:
str
-
n_rows
Data-row count, excluding the header.
- Type:
int
-
n_cols
Column count.
- Type:
int
-
file_bytes
File size in bytes.
- Type:
int
-
notes
Ingestion notes, for example when no dictionary sheet matched.
- Type:
str