Skip to content

fr_docket_links

Federal Register to docket links

The Federal Register ↔ docket bridge, from the Register's two statements: each federal_register row's printed docket_ids_json array exploded to one row per occurrence, and the docket of its regulations.gov link (regulations_dot_gov_docket_id) as a row of its own where no printed label carries it; link_source says which statement carries the pair. Printed labels are often agency docket numbers: of 2,791 documents that state a regulations.gov docket in windows sampled from 1998 to 2026, 1,513 print no label that carries it, and these WHERE docket_id = ? reaches only through the link; 998 print it in another form (Docket No. …), kept verbatim, with the id in normalized_docket_candidates_json (receipt round6/impl-W/m2/printed_vs_info.out). Carries the FR display columns. Use this instead of an ILIKE scan over federal_register.docket_ids_json: it is sorted by docket_id, so WHERE docket_id = ? prunes row groups. The rulemaking tables read the printed statement only (link_source printed or both). Built by build_fr_docket_links.

Coverage. Derived, and bounded by its inputs: one link per docket a dated Federal Register record names, in its printed labels or in its regulations.gov link, over the federal_register table's own date range. A Register row not read since the regulations.gov link was first requested contributes its printed labels only. (measured 2026-10-03)

  • Parquet file: fr_docket_links.parquet
  • MCP query_sql support: Configured; requires an available artifact.
  • Publication status: Not established by this schema page or its measurement date.
  • Row count: Not stated here; the MCP describe_table reply gives the live count under publication.
Column Type Description
docket_id VARCHAR The docket as the statement gives it: a printed label verbatim from docket_ids_json (Docket No. FAA-2007-29334, an agency docket number, a FRL- number), or the Register's regulations.gov docket id.
docket_source_ordinal BIGINT Zero-based position of a printed label in the original docket array, before null/blank elements are filtered; NULL on a row the regulations.gov link alone carries. With document_number and publication_date it identifies a printed occurrence.
normalized_docket_candidates_json VARCHAR Owner-normalizer candidate interpretations of the docket value; an empty array means no supported interpretation, not absence of a docket.
docket_normalization_rule VARCHAR Digest-pinned owner identifier-normalization rule used to produce candidates.
document_number VARCHAR Literal Federal Register document number. Join to federal_register on both this value and publication_date; a number can name distinct dated records.
title VARCHAR FR document title.
abstract VARCHAR FR document abstract.
document_type VARCHAR FR type: Rule, Proposed Rule, Notice, or Presidential Document.
subtype VARCHAR FR subtype where the Federal Register supplies one, e.g. Final Rule, Interim Final Rule.
publication_date VARCHAR Date the document published in the Federal Register; required with document_number for the FR record join.
effective_on VARCHAR Date the rule takes effect, when the document carries one.
comments_close_on VARCHAR Comment period close date, when the document carries one.
signing_date VARCHAR Signing date, for presidential documents.
agency_slugs VARCHAR Federal Register agency slugs for the issuing agencies.
docket_ids_json VARCHAR The full original docket-ID array, retained so consumers can see the other IDs this FR document links to.
regulation_id_numbers_json VARCHAR RINs (JSON array) claimed by the FR document — the RIN-side join key for a rulemaking.
html_url VARCHAR federalregister.gov HTML permalink.
pdf_url VARCHAR Federal Register PDF URL.
executive_order_number VARCHAR Executive order number, for presidential documents.
link_source VARCHAR Which of the Register's statements carries the pair: printed (the document's printed docket labels), regulations_dot_gov_info (the Register's regulations.gov link, read from regulations_dot_gov_docket_id alone, never off a document id) or both, where a printed label is that docket id or SpicyDocs' normalizer reads it from the label. Every printed occurrence that carries the pair reads both, and the link then adds no row. Never NULL.