comments_index
A tiny row-count rollup of comments per agency, docket, and posted month, so consumers can compute comment totals without scanning the full comments dataset. Rebuilt from the Iceberg catalog on every comments ETL run. It counts rows; it is not a list of files — the old comments/agency_code=/docket_id=/year=/month= tree it once described is no longer written, and building paths from this index 404s. Read comment rows from comments/agency/agency_code={X}/part-0.parquet instead.
- Parquet file:
comments_index.parquet - Queryable via MCP
query_sql: Yes
| Column | Type | Description |
|---|---|---|
agency_code |
VARCHAR |
Agency code of the counted comments. |
docket_id |
VARCHAR |
Docket id of the counted comments. |
year |
BIGINT |
Year of the comments' posted_date. |
month |
BIGINT |
Month (1–12) of the comments' posted_date. |
row_count |
BIGINT |
Number of comments in that agency / docket / month. |