scenario: us_mortgage
generated_from: dependency_graph.jsonl (audited, join-spine-repaired)
note: 'every join here was executed against duckdb: it returns rows. `join: via` means a two-hop path through `via.table`; use the `sql` field
  verbatim.'
n_joins: 2
joins:
- left_table: dim_geography
  left_col: county_code
  right_table: origination_hmda
  right_col: county_fips
  type: CROSSWALK_JOIN
  join: via
  via:
    table: xw_ny_county_fips
    left_col: county_code
    right_col: county_fips
    left_cov: 1.0
    right_cov: 1.0
    two_hop_left_cov: 1.0
  strength: high
  key_coverage: 0.0
  sql: "SELECT * FROM \"dim_geography\" a\n  JOIN \"xw_ny_county_fips\" x ON CAST(a.\"county_code\" AS VARCHAR) = CAST(x.\"county_code\" AS VARCHAR)\n\
    \  JOIN \"origination_hmda\" b ON CAST(x.\"county_fips\" AS VARCHAR) = CAST(b.\"county_fips\" AS VARCHAR)"
- left_table: origination_hmda
  left_col: county_name
  right_table: dim_geography
  right_col: county_code
  type: CROSSWALK_JOIN
  join: via
  via:
    table: xw_ny_county_fips
    left_col: county_name
    right_col: county_code
    left_cov: 0.9841
    right_cov: 1.0
    two_hop_left_cov: 0.9841
  strength: medium
  key_coverage: 0.0
  sql: "SELECT * FROM \"origination_hmda\" a\n  JOIN \"xw_ny_county_fips\" x ON CAST(a.\"county_name\" AS VARCHAR) = CAST(x.\"county_name\" AS\
    \ VARCHAR)\n  JOIN \"dim_geography\" b ON CAST(x.\"county_code\" AS VARCHAR) = CAST(b.\"county_code\" AS VARCHAR)"