# Vlada benefit reading, each contract's closest contracts by footprint (edition v20261018b)

`medicare_semantic.vw_benefit_contract_match` · Medicare Advantage, Part D, Medicaid · status live · 15,180 rows

For each contract, product type and plan type: the other contracts with the same product and plan type and shared counties, ranked by the member-weighted county Jaccard (SUM(min) / SUM(max)); floors: 1,000 shared-county members on side A and a Jaccard of 0.05. overlap_share_a / _b = each side's members in the shared counties. Pointer -> medicare_semantic.benefit_contract_match_v20261018b. Owner: the Benefits lane.

**Grain.** One row per base_year x bucket x plan_type x contract_id x match_contract_id (top 10).

| column | type | meaning |
|---|---|---|
| base_year | int |  |
| bucket | string |  |
| plan_type | string |  |
| contract_id | string |  |
| parent_a | string |  |
| match_contract_id | string |  |
| parent_b | string |  |
| members_a | double |  |
| members_b | double |  |
| counties_a | bigint |  |
| counties_b | bigint |  |
| shared_counties | bigint |  |
| shared_a | double |  |
| shared_b | double |  |
| overlap_share_a | double |  |
| overlap_share_b | double |  |
| jaccard | double |  |
| match_rank | bigint |  |

Tools that read it: get_benefit_change

Query it: `query_substrate` with `SELECT ... FROM medicare_semantic.vw_benefit_contract_match WHERE ... LIMIT 100`; read `describe_data_source("medicare_semantic.vw_benefit_contract_match")` first.