section_diff_items
Bill comparison items
One row per settled correspondence in one version-pair comparison. text_diff_json is byte-capped, and text_diff_truncated says when the cap was reached. All columns are stored as VARCHAR.
Coverage. Sampled: the settled correspondences within each comparison in section_diffs, so it covers exactly what that table covers. (measured 2026-09-28)
- Parquet file:
section_diff_items.parquet - MCP
query_sqlsupport: 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_tablereply gives the live count underpublication.
| Column | Type | Description |
|---|---|---|
bill_id |
VARCHAR |
The bill both printings belong to. |
from_version_code |
VARCHAR |
The earlier printing's version code. |
from_source |
VARCHAR |
Which acquisition path supplied the earlier printing. |
to_version_code |
VARCHAR |
The later printing's version code. |
to_source |
VARCHAR |
Which acquisition path supplied the later printing. |
seq |
VARCHAR |
Position in the engine's own stable ordering of settled correspondences. |
op |
VARCHAR |
What changed: added, removed, modified, unchanged or moved. |
similarity |
VARCHAR |
The word-overlap signal the assignment round recorded, rounded to four places. |
moved |
VARCHAR |
True when the engine classified this correspondence as a move. |
pairing_rule |
VARCHAR |
Which assignment round selected this correspondence: path-round, move-round or unpaired. |
evidence_json |
VARCHAR |
Every named signal the round recorded, as the engine spelled them. |
match_path |
VARCHAR |
The normalized cross-version key, unit-separator joined. |
match_path_json |
VARCHAR |
The same path as a JSON array. |
display_path_old_json |
VARCHAR |
The earlier printing's display path, as a JSON array. |
display_path_new_json |
VARCHAR |
The later printing's display path, as a JSON array. |
section_number |
VARCHAR |
The section's displayed number. |
heading |
VARCHAR |
The section heading, preferring the later printing's. |
from_element_id |
VARCHAR |
The earlier printing's own element id, which resolves this side to a bill_sections row. |
to_element_id |
VARCHAR |
The later printing's own element id. |
from_text_sha256 |
VARCHAR |
Digest of the earlier body text, so an unchanged side is recognisable without it. |
to_text_sha256 |
VARCHAR |
Digest of the later body text. |
from_text_chars |
VARCHAR |
Character length of the earlier body text. |
to_text_chars |
VARCHAR |
Character length of the later body text. |
text_diff_json |
VARCHAR |
The engine's word-level diff as a JSON array, capped at the published byte limit. |
text_diff_truncated |
VARCHAR |
True when the word diff did not fit the cap and entries were dropped from the end. |
financial_from_amounts_json |
VARCHAR |
Every dollar figure the earlier body states, as a JSON array. |
financial_to_amounts_json |
VARCHAR |
Every dollar figure the later body states, as a JSON array. |
financial_amounts_changed |
VARCHAR |
True when the two multisets of figures differ; needs no account model to be true. |
financial_has_amendment_annotations |
VARCHAR |
True when the section carries amendment annotations around its figures. |